Professional Documents
Culture Documents
Heat Exchangers: Design, Operation, Maintenance and Enhancement
Heat Exchangers: Design, Operation, Maintenance and Enhancement
o
a
t
i
n
g
h
e
a
d
p
a
c
k
e
d
P
u
l
l
-
t
h
r
o
u
g
h
o
f
d
e
s
i
g
n
t
u
b
e
s
h
e
e
t
U
-
t
u
b
e
o
a
t
i
n
g
h
e
a
d
(
s
p
l
i
t
b
a
c
k
i
n
g
r
i
n
g
)
o
a
t
i
n
g
h
e
a
d
o
a
t
i
n
g
h
e
a
d
T
.
E
.
M
.
A
.
r
e
a
r
-
h
e
a
d
t
y
p
e
L
o
r
M
o
r
N
U
W
S
P
T
R
e
l
a
t
i
v
e
c
o
s
t
i
n
c
r
e
a
s
e
s
f
r
o
m
A
(
l
e
a
s
t
e
x
p
e
n
s
i
v
e
)
t
h
r
o
u
g
h
E
(
m
o
s
t
e
x
p
e
n
s
i
v
e
)
B
A
C
E
D
E
P
r
o
v
i
s
i
o
n
f
o
r
d
i
e
r
e
n
t
i
a
l
e
x
p
a
n
s
i
o
n
E
x
p
a
n
s
i
o
n
I
n
d
i
v
i
d
u
a
l
t
u
b
e
s
F
l
o
a
t
i
n
g
h
e
a
d
F
l
o
a
t
i
n
g
h
e
a
d
F
l
o
a
t
i
n
g
h
e
a
d
F
l
o
a
t
i
n
g
h
e
a
d
j
o
i
n
t
i
n
f
r
e
e
t
o
e
x
p
a
n
d
s
h
e
l
l
R
e
m
o
v
a
b
l
e
b
u
n
d
l
e
N
o
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
R
e
p
l
a
c
e
m
e
n
t
b
u
n
d
l
e
p
o
s
s
i
b
l
e
N
o
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
I
n
d
i
v
i
d
u
a
l
t
u
b
e
s
r
e
p
l
a
c
e
a
b
l
e
Y
e
s
O
n
l
y
t
h
o
s
e
i
n
Y
e
s
Y
e
s
Y
e
s
Y
e
s
o
u
t
s
i
d
e
r
o
w
T
u
b
e
c
l
e
a
n
i
n
g
b
y
c
h
e
m
i
c
a
l
s
i
n
s
i
d
e
a
n
d
o
u
t
s
i
d
e
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
I
n
t
e
r
i
o
r
t
u
b
e
c
l
e
a
n
i
n
g
m
e
c
h
a
n
i
c
a
l
l
y
Y
e
s
S
p
e
c
i
a
l
t
o
o
l
s
r
e
q
u
i
r
e
d
Y
e
s
Y
e
s
Y
e
s
Y
e
s
E
x
t
e
r
i
o
r
t
u
b
e
c
l
e
a
n
i
n
g
m
e
c
h
a
n
i
c
a
l
l
y
:
T
r
i
a
n
g
u
l
a
r
p
i
t
c
h
N
o
N
o
N
o
N
o
N
o
N
o
S
q
u
a
r
e
p
i
t
c
h
N
o
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
H
y
d
r
a
u
l
i
c
-
j
e
t
c
l
e
a
n
i
n
g
:
T
u
b
e
i
n
t
e
r
i
o
r
Y
e
s
S
p
e
c
i
a
l
t
o
o
l
s
r
e
q
u
i
r
e
d
Y
e
s
Y
e
s
Y
e
s
Y
e
s
T
u
b
e
e
x
t
e
r
i
o
r
N
o
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
D
o
u
b
l
e
t
u
b
e
s
h
e
e
t
f
e
a
s
i
b
l
e
Y
e
s
Y
e
s
N
o
N
o
Y
e
s
N
o
N
u
m
b
e
r
o
f
t
u
b
e
p
a
s
s
e
s
N
o
p
r
a
c
t
i
c
a
l
A
n
y
e
v
e
n
L
i
m
i
t
e
d
t
o
o
n
e
N
o
p
r
a
c
t
i
c
a
l
N
o
p
r
a
c
t
i
c
a
l
N
o
p
r
a
c
t
i
c
a
l
l
i
m
i
t
a
t
i
o
n
s
n
u
m
b
e
r
p
o
s
s
i
b
l
e
o
r
t
w
o
p
a
s
s
e
s
l
i
m
i
t
a
t
i
o
n
s
l
i
m
i
t
a
t
i
o
n
s
l
i
m
i
t
a
t
i
o
n
s
I
n
t
e
r
n
a
l
g
a
s
k
e
t
s
e
l
i
m
i
n
a
t
e
d
Y
e
s
Y
e
s
Y
e
s
N
o
Y
e
s
N
o
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.2 Classication by construction STHE 33
3.2 Classication by construction STHE
Fig. 3.2 to Fig. 3.8 show details of the construction of the TEMA types of shell-and-tube
heat exchangers. These types are:
Fixed tube sheet
U-tube
Floating head
3.2.1 Fixed tube sheet
Fixed-tube-sheet exchangers (Fig. 3.2) are used more often than any other type, and
the frequency of use has been increasing in recent years. The tube sheets are welded
to the shell. Usually these extend beyond the shell and serve as anges to which the
tube-side headers are bolted. This construction requires that the shell and tube-sheet
materials be weldable to each other. When such welding is not possible, a blind-gasket
type of construction is utilized. The blind gasket is not accessible for maintenance or
replacement once the unit has been constructed. This construction is used for steam
surface condensers, which operate under vacuum.
Figure 3.2. Heat-exchanger-component nomenclature. Fixed tube heat sheet shell and tube
heat exchanger. (Standard of Tubular Exchanger Manufacturers Association, 6th ed., 1978.)
The tube-side header (or channel) may be welded to the tube sheet, as shown in Fig. 3.1
for type C and N heads. This type of construction is less costly than types B and M or
A and L and still oers the advantage that tubes may be examined and replaced without
disturbing the tube-side piping connections. There is no limitation on the number of
tube-side passes. Shell-side passes can be one or more, although shells with more than
two shell side passes are rarely used. Tubes can completely ll the heat-exchanger shell.
Clearance between the outermost tubes and the shell is only the minimum necessary
for fabrication. Between the inside of the shell and the baes some clearance must be
provided so that baes can slide into the shell. Fabrication tolerances then require some
additional clearance between the outside of the baes and the outermost tubes. The edge
distance between the outer tube limit (OTL) and the bae diameter must be sucient
to prevent vibration of the tubes from breaking through the bae holes. The outermost
tube must be contained within the OTL.
Clearances between the inside shell diameter and OTL are 13 mm (1/2 in) for 635-mm-
(25-in-) inside-diameter shells and up, 11 mm for 254- through 610-mm (10- through
24-in) pipe shells, and slightly less for smaller-diameter pipe shells.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
34 3 Code and standards
Tubes can be replaced. Tube-side headers, channel covers, gaskets, etc., are accessible for
maintenance and replacement. Neither the shell-side bae structure nor the blind gasket
is accessible. During tube removal, a tube may break within the shell. When this occurs,
it is most dicult to remove or to replace the tube. The usual procedure is to plug the
appropriate holes in the tube sheets.
Dierential expansion between the shell and the tubes can develop because of dierences
in length caused by thermal expansion. Various types of expansion joints are used to
eliminate excessive stresses caused by expansion. The need for an expansion joint is a
function of both the amount of dierential expansion and the cycling conditions to be
expected during operation. A number of types of expansion joints are available (Fig. 3.3)
Figure 3.3. Expansion joints.
.
a Flat plates. Two concentric at plates with a bar at the outer edges. The at plates
can ex to make some allowance for dierential expansion. This design is generally
used for vacuum service and gauge pressures below 103 kPa (15 lbf/in2). All welds
are subject to severe stress during dierential expansion.
b Flanged-only heads. The at plates are anged (or curved). The diameter of these
heads is generally 203 mm (8 in) or more greater than the shell diameter. The
welded joint at the shell is subject to the stress referred to before, but the joint
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.2 Classication by construction STHE 35
connecting the heads is subjected to less stress during expansion because of the
curved shape.
c Flared shell or pipe segments. The shell may be ared to connect with a pipe
section, or a pipe may be halved and quartered to produce a ring.
d Formed heads. A pair of dished-only or elliptical or anged and dished heads can
be used. These are welded together or connected by a ring. This type of joint is
similar to the anged-only-head type but apparently is subject to less stress.
e Flanged and ued heads. A pair of anged-only heads is provided with concentric
reverse ue holes. These heads are relatively expensive because of the cost of the
uing operation. The curved shape of the heads reduces the amount of stress at the
welds to the shell and also connecting the heads.
f Toroidal. The toroidal joint has a mathematically predictable smooth stress pat-
tern of low magnitude, with maximum stresses at sidewalls of the corrugation and
minimum stresses at top and bottom. The foregoing designs were discussed as ring
expansion joints by Kopp and Sayre, Expansion Joints for Heat Exchangers (ASME
Misc. Pap., vol. 6, no. 211). All are statically indeterminate but are subjected
to analysis by introducing various simplifying assumptions. Some joints in current
industrial use are of lighter wall construction than is indicated by the method of
this paper.
g Bellows. Thin-wall bellows joints are produced by various manufacturers. These are
designed for dierential expansion and are tested for axial and transverse movement
as well as for cyclical life. Bellows may be of stainless steel, nickel alloys, or copper.
(Aluminum, Monel, phosphor bronze, and titanium bellows have been manufac-
tured.) Welding nipples of the same composition as the heat-exchanger shell are
generally furnished. The bellows may be hydraulically formed from a single piece
of metal or may consist of welded pieces. External insulation covers of carbon steel
are often provided to protect the light-gauge bellows from damage. The cover also
prevents insulation from interfering with movement of the bellows (see h).
h Toroidal bellows. For high-pressure service the bellows type of joint has been modi-
ed so that movement is taken up by thin-wall small-diameter bellows of a toroidal
shape. Thickness of parts under high pressure is reduced considerably (see f ).
Improper handling during manufacture, transit, installation, or maintenance of the heat
exchanger equipped with the thin-wallbellows type or toroidal type of expansion joint can
damage the joint. In larger units these light-wall joints are particularly susceptible to
damage, and some designers prefer the use of the heavier walls of formed heads.
Chemical-plant exchangers requiring expansion joints most commonly have used the
anged-and-ued-head type. There is a trend toward more common use of the light-
wall-bellows type.
3.2.2 U-Tube Heat Exchanger
Fig. 3.4 shows U-tube heat exchanger Type CFU. The tube bundle consists of a stationary
tube sheet, U tubes (or hairpin tubes), baes or support plates, and appropriate tie rods
and spacers. The tube bundle can be removed from the heat-exchanger shell. A tube-side
header (stationary head) and a shell with integral shell cover, which is welded to the
shell, are provided. Each tube is free to expand or contract without any limitation being
placed upon it by the other tubes. The U-tube bundle has the advantage of providing
minimum clearance between the outer tube limit and the inside of the shell for any of
the removable-tube-bundle constructions. Clearances are of the same magnitude as for
xed-tube-sheet heat exchangers. The number of tube holes in a given shell is less than
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
36 3 Code and standards
that for a xed-tube-sheet exchanger because of limitations on bending tubes of a very
short radius.
Figure 3.4. Heat-exchanger-component nomenclature. U-tube heat exchanger. Type CFU.
(Standard of Tubular Exchanger Manufacturers Association, 6th ed., 1978.)
The U-tube design oers the advantage of reducing the number of joints. In high-pressure
construction this feature becomes of considerable importance in reducing both initial and
maintenance costs. The use of U-tube construction has increased signicantly with the
development of hydraulic tube cleaners, which can remove fouling residues from both the
straight and the U-bend portions of the tubes. Rods and conventional mechanical tube
cleaners cannot pass from one end of the U tube to the other. Power-driven tube cleaners,
which can clean both the straight legs of the tubes and the bends, are available. Hydraulic
jetting with water forced through spray nozzles at high pressure for cleaning tube interiors
and exteriors of removal bundles is reported in the recent ASME publications.
U-tube can be used for high pressure and high temperature application like kettle reboiler,
evaporator, tank section heaters ,etc.
The tank suction heater, as illustrated in Fig. 3.5, contains a U-tube bundle. This design
is often used with outdoor storage tanks for heavy fuel oils, tar, molasses, and similar
uids whose viscosity must be lowered to permit easy pumping. Uusally the tube-side
heating medium is steam. One end of the heater shell is open, and the liquid being heated
passes across the outside of the tubes. Pumping costs can be reduced without heating the
entire contents of the tank. Bare tube and integral low-n tubes are provided with baes.
Longitudinal n-tube heaters are not baed. Fins are most often used to minimize the
fouling potential in these uids.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.2 Classication by construction STHE 37
Figure 3.5. Heat-exchanger-component nomenclature. U-tube heat exchanger. Type CFU.
(Standard of Tubular Exchanger Manufacturers Association, 6th ed., 1978.)
Kettle-type reboilers, evaporators, etc. , are often U-tube exchangers with enlarged shell
sections for vapor-liquid separation (Fig.3.6). The U-tube bundle replaces the oating-
heat bundle of Fig. 3.4.
Figure 3.6. Kettle reboiler
The U-tube exchanger with copper tubes, cast-iron header, and other parts of carbon
steel is used for water and steam services in oce buildings, schools, hospitals, hotels, etc.
Nonferrous tube sheets and admiralty or 90-10 copper-nickel tubes are the most frequently
used substitute materials. These standard exchangers are available from a number of
manufacturers at costs far below those of custombuilt process-industry equipment.
3.2.3 Floating Head Designs
In an eort to reduce thermal stresses and provide a means to remove the tube bundle
for cleaning, several oating rear head designs have been established. The simplest is a
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
38 3 Code and standards
Internal oating head (pull- through design) Fig3.9 design which allows the tube bundle to
be pulled entirely through the shell for service or replacement. In order to accommodate
the rear head bolt circle, tubes must be removed resulting in a less ecient use of shell
size. In addition, the missing tubes result in larger annular spaces and can contribute to
reduced ow across the eective tube surface, resulting in reduced thermal performance.
Some designs include sealing strips installed in the shell to help block the bypass steam.
Another oating head design that partially addresses the above disadvantages is a split-
ring oating head. Here the oating head bonnet is bolted to a split backing ring instead
of the tube sheet. This eliminates the bolt circle diameter and allows a full complement
of tubes to ll the shell. This construction is more expensive than a common pull through
design, but is in wide use in petrochemical applications. For applications with high
pressures or temperatures, or where more positive sealing between the uids is desired,
the pull-through design should be specied.
Two other types, the outside packed lantern ring and the outside packed stung box
designs oer less positive sealing against leakage to the atmosphere than the pull though
or split ring designs, but can be congured for single tube pass duty. More details about
the various types of oating head shell and tube heat exchanger is given the following
sections
Packed-Lantern-Ring Exchanger: (Fig. 3.7 ) This construction is the least costly
of the straight-tube removable bundle types. The shell- and tube-side uids are each
contained by separate rings of packing separated by a lantern ring and are installed at the
oating tube sheet. The lantern ring is provided with weep holes. Any leakage passing
the packing goes through the weep holes and then drops to the ground. Leakage at the
packing will not result in mixing within the exchanger of the two uids. The width of the
oating tube sheet must be great enough to allow for the packings, the lantern ring, and
dierential expansion. Sometimes a small skirt is attached to a thin tube sheet to provide
the required bearing surface for packings and lantern ring. The clearance between the
outer tube limit and the inside of the shell is slightly larger than that for xed-tube-sheet
and U-tube exchangers.
The use of a oating-tube-sheet skirt increases this clearance. Without the skirt the
clearance must make allowance for tubehole distortion during tube rolling near the outside
edge of the tube sheet or for tube-end welding at the oating tube sheet.
The packed-lantern-ring construction is generally limited to design temperatures below
191
C (375
F) and to the mild services of water, steam, air, lubricating oil, etc. Design
gauge pressure does not exceed 2068 kPa (300 lbf/in
2
) for pipe shell exchangers and is
limited to 1034 kPa (150 lbf/in
2
) for 610- to 1067-mm- (24- to 42-in-) diameter shells.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.2 Classication by construction STHE 39
Figure 3.7. Heat-exchanger-component nomenclature. Exchanger with packed oating tube
sheet and lantern ring. Type AJW. External oating head design. (Standard of Tubular Ex-
changer Manufacturers Association, 6th ed., 1978.)
Outside-Packed Floating-Head Exchanger: (Fig. 3.8) The shell-side uid is con-
tained by rings of packing, which are compressed within a stung box by a packing
follower ring. This construction was frequently used in the chemical industry, but in
recent years usage has decreased. The removable-bundle construction accommodates dif-
ferential expansion between shell and tubes and is used for shell-side service up to 4137
kPa gauge pressure (600 lbf/in2) at 316
C (600
F).
Figure 3.8. Heat-exchanger-component nomenclature. Outside-packed oating-head ex-
changer. Type AEP. (Standard of Tubular Exchanger Manufacturers Association, 6th ed., 1978.)
There are no limitations upon the number of tube-side passes or upon the tube-side
design pressure and temperature. The outside-packed oating-head exchanger was the
most commonly used type of removable- bundle construction in chemical-plant service.
The oating-tube-sheet skirt, where in contact with the rings of packing, has ne machine
nish. A split shear ring is inserted into a groove in the oating-tube-sheet skirt. A slip-
on backing ange, which in service is held in place by the shear ring, bolts to the external
oating- head cover. The oating-head cover is usually a circular disk. With an odd
number of tube-side passes, an axial nozzle can be installed in such a oating- head cover.
If a side nozzle is required, the circular disk is replaced by either a dished head or a channel
barrel (similar to Fig. 11-36f ) bolted between oating-head cover and oating-tube-sheet
skirt. The outer tube limit approaches the inside of the skirt but is farther removed from
the inside of the shell than for any of the previously discussed constructions. Clearances
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
40 3 Code and standards
between shell diameter and bundle OTL are 22 mm (7.8 in) for small-diameter pipe shells,
44 mm (1e in) for large-diameter pipe shells, and 58 mm (2g in) for moderatediameter
plate shells.
Internal Floating-Head Exchanger: (Fig. 3.9) The internal oating-head design
is used extensively in petroleum-renery service, but in recent years there has been a
decline in usage. The tube bundle is removable, and the oating tube sheet moves (or
oats) to accommodate dierential expansion between shell and tubes. The outer tube
limit approaches the inside diameter of the gasket at the oating tube sheet. Clearances
(between shell and OTL) are 29 mm for pipe shells and 37 mm for moderatediameter plate
shells. A split backing ring and bolting usually hold the oating-head cover at the oating
tube sheet. These are located beyond the end of the shell and within the larger-diameter
shell cover. Shell cover, split backing ring, and oating-head cover must be removed before
the tube bundle can pass through the exchanger shell. With an even number of tube-side
passes the oating-head cover serves as return cover for the tube-side uid. With an odd
number of passes a nozzle pipe must extend from the oating-head cover through the shell
cover. Provision for both dierential expansion and tube-bundle removal must be made.
Figure 3.9. Heat-exchanger-component nomenclature. Internal oating head (pull- through
design). Type AES. (Standard of Tubular Exchanger Manufacturers Association, 6th ed., 1978.)
Figure 3.10. Heat-exchanger-component nomenclature. Exchanger with packed oating tube
sheet and lantern ring. Type AES. (Standard of Tubular Exchanger Manufacturers Association,
6th ed., 1978.)
Pull-Through Floating-Head Exchanger: (Fig. 3.12) Construction is similar to that
of the internal-oating-head split-backing ring exchanger except that the oating-head
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.3 Shell Constructions 41
cover bolts directly to the oating tube sheet. The tube bundle can be withdrawn from
the shell without removing either shell cover or oating-head cover. This feature reduces
maintenance time during inspection and repair.
The large clearance between the tubes and the shell must provide for both the gasket
and the bolting at the oating-head cover. This clearance is about 2 to 2.5 times that
required by the split-ring design. Sealing strips or dummy tubes are often installed to
reduce bypassing of the tube bundle.
Figure 3.11. Heat-exchanger-component nomenclature. Kettle-type oating-head reboiler.
Type AKT. (Standard of Tubular Exchanger Manufacturers Association, 6th ed., 1978.)
3.3 Shell Constructions
The most common TEMA shell type is the E shell as it is most suitable for most
industrial process cooling applications. However, for certain applications, other
shells oer distinct advantages. For example, the TEMA-F shell design provides
for a longitudinal ow plate to be installed inside the tube bundle assembly. This
plate causes the shell uid to travel down one half of the tube bundle, then down
the other half, in eect producing a counter-current ow pattern which is best for
heat transfer. This type of construction can be specied where a close approach
temperature is required and when the ow rate permits the use of one half of the
shell at a time. In heat recovery applications, or where the application calls for
increased thermal length to achieve eective overall heat transfer, shells can be
installed with the ows in series. Up to six shorter shells in series is common and
results in counter-current ow close to performance as if one long shell in a single
pass design were used.
TEMA G and H shell designs are most suitable for phase change applications where
the bypass around the longitudinal plate and counter-current ow is less impor-
tant than even ow distribution. In this type of shell, the longitudinal plate oers
better ow distribution in vapor streams and helps to ush out non-condensable.
They are frequently specied for use in horizontal thermosiphon reboilers and total
condensers.
TEMA J Shells are typically specied for phase change duties where signicantly
reduced shell side pressure drops are required. They are commonly used in stacked
sets with the single nozzles used as the inlet and outlet. A special type of J-shell
is used for ooded evaporation of shell side uids. A separate vapor disengagement
vessel without tubes is installed above the main J shell with the vapor outlet at the
top of this vessel. The
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
42 3 Code and standards
TEMA K shell, also termed a kettle reboiler, is specied when the shell side stream
will undergo vaporization. The liquid level of a K shell design should just cover the
tube bundle, which lls the smaller diameter end of the shell. This liquid level is
controlled by the liquid owing over a weir at the far end of the entrance nozzle. The
expanded shell area serves to facilitate vapor disengagement for boiling liquid in the
bottom of the shell. To insure against excessive liquid carry-though with the vapor
stream, a separate vessel as described above is specied. Liquid carry-through can
also be minimized by installing a mesh demister at the vapor exit nozzle. U-bundles
are typically used with K shell designs. K shells are expensive for high pressure
vaporization due to shell diameter and the required wall thickness.
The TEMA X shell, or crossow shell is most commonly used in vapor condensing
applications, though it can also be used eectively in low pressure gas cooling or
heating. It produces a very low shell side pressure drop, and is therefore most
suitable for vacuum service condensing. In order to assure adequate distribution
of vapors, X-shell designs typically feature an area free of tubes along the top of
the exchanger. It is also typical to design X shell condensers with a ow area at
the bottom of the tube bundle to allow free condensate ow to the exit nozzle.
Careful attention to the eective removal of non-condensables is vital to X-shell
constructions.
3.4 Tube side construction
3.4.1 Tube-Side Header:
The tube-side header (or stationary head) contains one or more ow nozzles.
The bonnet (Fig. 3.1B) bolts to the shell. It is necessary to remove the bonnet in
order to examine the tube ends. The xed-tubesheet exchanger of Fig. 3.1b has
bonnets at both ends of the shell.
The channel (Fig. 3.1A) has a removable channel cover. The tube ends can be
examined by removing this cover without disturbing the piping connections to the
channel nozzles. The channel can bolt to the shell as shown in Fig. 3.1a and c.
The Type C and Type N channels of Fig. 3.1 are welded to the tube sheet. This
design is comparable in cost with the bonnet but has the advantages of permitting
access to the tubes without disturbing the piping connections and of eliminating a
gasketed joint.
Special High-Pressure Closures (Fig. 3.1D) The channel barrel and the tube sheet
are generally forged. The removable channel cover is seated in place by hydrostatic
pressure, while a shear ring subjected to shearing stress absorbs the end force. For
pressures above 6205 kPa (900 lbf/in2) these designs are generally more economical
than bolted constructions, which require larger anges and bolting as pressure in-
creases in order to contain the end force with bolts in tension. Relatively light-gauge
internal pass partitions are provided to direct the ow of tube-side uids but are
designed only for the dierential pressure across the tube bundle.
3.4.2 Tube-Side Passes
Most exchangers have an even number of tube-side passes. The xed-tube-sheet exchanger
(which has no shell cover) usually has a return cover without any ow nozzles as shown in
Fig. 3.1M; Types L and N are also used. All removable-bundle designs (except for the U
tube) have a oating-head cover directing the ow of tube-side uid at the oating tube
sheet.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.4 Tube side construction 43
3.4.3 Tubes Type
There are dierent type of tubes used in heat exchangers. These are
1. Plain tube
(a) Straight tube
(b) U-tube with a U-bend
(c) Coiled tubes
2. Finned tube
3. Duplex or bimetallic tube. These tube are in reality two tube of dierent materials,
one closely tted over the other with no gap between them. They are made by
drawing the outer tube onto the inner one or by shrink tting. These are used
where corrosive nature of the tube side uid is such that no one metal or alloy is
compatible with uids.
4. Enhanced surface tube
1. Plain tube
Standard heat-exchanger tubing is (1/4, 3/8, 1/2, 5/8, 3/4, 1, 1 1/4, 1 1/2 inch in
outside diameter (1 inch= 25.4 mm). Wall thickness is measured in Birmingham
wire gauge (BWG) units. The most commonly used tubes in chemical plants and
petroleum reneries are 19- and 25-mm (3/4- and 1-in) outside diameter. Standard
tube lengths are 8, 10, 12, 16, and 20 ft, with 20 ft now the most common ( 1 ft=
0.3048 m).
Manufacturing tolerances for steel, stainless-steel, and nickel alloy tubes are such
that the tubing is produced to either average or minimum wall thickness. Seamless
carbon steel tube of minimum wall thickness may vary from 0 to 20 percent above the
nominal wall thickness. Average-wall seamless tubing has an allowable variation of
plus or minus 10 percent. Welded carbon steel tube is produced to closer tolerances
(0 to plus 18 percent on minimum wall; plus or minus 9 percent on average wall).
Tubing of aluminum, copper, and their alloys can be drawn easily and usually is
made to minimum wall specications.
Common practice is to specify exchanger surface in terms of total external square
feet of tubing. The eective outside heat-transfer surface is based on the length of
tubes measured between the inner faces of tube sheets. In most heat exchangers
there is little dierence between the total and the eective surface. Signicant
dierences are usually found in high-pressure and double-tube-sheet designs.
Tube thickness The tube should be able to stand:
(a) pressure on the inside and out side of the tube
(b) temperature on both the sides
(c) thermal stress due to the dierential expansion of the shell and the tube bundle
(d) corrosive nature of both the shell-side and the tube side uid
The tube thickness is given a function of the tube out side diameter in accordance
with B.W.G.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
44 3 Code and standards
Figure 3.12. Tube thickness
2. Finned tube: As the name implies, nned tube have ns to the tubular surface.
Fins can be longtiudinal, radial or helical and may be on the outside or inside or on
both sides of the tube. Fig. 5.7shows some of the commonly used ns. The ns are
generally used when at least one of the uid is gas.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.4 Tube side construction 45
Figure 3.13. Examples of extended surfaces on one or both sides. (a) Radial ns. (b) Serrated
radial ns. (c) Studded surface. (d) Joint between tubesheet and low n tube with three times
bare surface. (e) External axial ns. ( f ) Internal axial ns. (9) Finned surface with internal
spiral to promote turbulence. (h) Plate ns on both sides. (i) Tubes and plate ns.
(a) Integrally nned tube, which is available in a variety of alloys and sizes, is
being used in shell-and-tube heat exchangers. The ns are radially extruded
from thick-walled tube to a height of 1.6 mm (1/16 in) spaced at 1.33 mm (19
ns per inch) or to a height of 3.2 mm (1/8 in) spaced at 2.3 mm (11 ns per
inch). External surface is approximately 2 1/2 times the outside surface of a
bare tube with the same outside diameter. Also available are 0.93-mm- (0.037-
in-) high ns spaced 0.91 mm (28 ns per inch) with an external surface about
3.5 times the surface of the bare tube. Bare ends of nominal tube diameter are
provided, while the n height is slightly less than this diameter. The tube can
be inserted into a conventional tube bundle and rolled or welded to the tube
sheet by the same means, used for bare tubes. An integrally nned tube rolled
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
46 3 Code and standards
into a tube sheet with double serrations and ared at the inlet is shown in
Fig. 11-39. Internally nned tubes have been manufactured but have limited
application.
(b) Longitudinal ns are commonly used in double-pipe exchangers upon the
outside of the inner tube. U-tube and conventional removable tube bundles
are also made from such tubing. The ratio of external to internal surface
generally is about 10 or 15:1.
(c) Transverse ns upon tubes are used in low-pressure gas services. The primary
application is in air-cooled heat exchangers (as discussed under that heading),
but shell-and-tube exchangers with these tubes are in service.
3. Bimetallic Tubes When corrosive requirements or temperature conditions do not
permit the use of a single alloy for the tubes, bimetallic (or duplex) tubes may be
used. These can be made from almost any possible combination of metals. Tube
sizes and gauges can be varied. For thin gauges the wall thickness is generally
divided equally between the two components. In heavier gauges the more expensive
component may comprise from a fth to a third of the total thickness.
The component materials comply with applicable ASTM specications, but after
manufacture the outer component may increase in hardness beyond specication
limits, and special care is required during the tube-rolling operation. When the
harder material is on the outside, precautions must be exercised to expand the
tube properly. When the inner material is considerably softer, rolling may not be
practical unless ferrules of the soft material are used.
In order to eliminate galvanic action the outer tube material may be stripped from
the tube ends and replaced with ferrules of the inner tube material. When the end
of a tube with a ferrule is expanded or welded to a tube sheet, the tube-side uid
can contact only the inner tube material, while the outer material is exposed to the
shell-side uid. Bimetallic tubes are available from a small number of tube mills
and are manufactured only on special order and in large quantities.
4. Enhance surface These kind of tubes enhance the heat transfer coecient (Fig.
5.7h,i). This may be achieved by two techniques.
(a) The surface is contoured or grooved in a variety of ways forming valley and
ridges. These are applicable in condenser and.
(b) The surface is prepared with special coating to provide a large number of
nucleation sites for use in boiling operations.
3.4.4 Tube arrangement
The tubes in an exchanger are usually arranged in an equilateral triangular, aquare or
rotated square pattern see g.3.14.
The triangular and rotated square pattern give higher heat transfer rates, but at the
expenses of higher pressure drop than the the square pattern. Square or rotated square
are used for hihger fouling uid, where it is necessary to mechanically clean the outside
of the tubes. The recommend tube pitch is P
t
= 1.25d
o
. Where square pattern is used
for easer of cleaning, the recommended minimum clearance between the tubes is 0.25 in
(6.4 mm)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.5 Shell side construction 47
p
t
d
o
Square pitch
pt
Equilateral triangular pitch
p
t
d
o
Rotaed square
Flow
Figure 3.14. Tube patterns.
3.4.5 Tube side passes
The uid in the tube is usually directed to ow back and forth in a number of passes
through groups of tube arranged in parallel to increase the length of the ow path. The
number of passes is selected to give the required side design velocity. Exchangers are built
form one to up to 16 passes. The tube are arranged into the number of passes required by
dividing up the exchanger headers (channels) with partition plates (pass partition) The
arrangement of the pass partition for 2,4 and 6 are shown in g.3.19
1
2
1
2
3
4
1
2
3
4 5
6
Two tube passes
Four tube passes
Six tube passes
1
2
3
4 5
6
Figure 3.15. Tube arrangement: showing pass-partitions in headers.
3.5 Shell side construction
3.5.1 Shell Sizes
Heat-exchanger shells are generally made from standard- wall steel pipe in sizes up to
305-mm (12-in) diameter; from 9.5-mm (3/8 in) wall pipe in sizes from 356 to 610 mm
(14 to 24 in); and from steel plate rolled at discrete intervals in larger sizes. Clearances
between the outer tube limit and the shell are discussed elsewhere in connection with the
dierent types of construction.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
48 3 Code and standards
3.5.2 Shell-Side Arrangements
1. The one-pass shell (Fig. 3.1E) is the most commonly used arrangement. Con-
densers from single component vapors often have the nozzles moved to the center
of the shell for vacuum and steam services. Solid longitudinal bae is provided to
form a two-pass shell (Fig. 3.1F). It may be insulated to improve thermal eciency.
(See further discussion on baes).
2. A two-pass shell can improve thermal eectiveness at a cost lower than for two
shells in series.
3. For split ow (Fig. 3.1G), the longitudinal bae may be solid or perforated. The
latter feature is used with condensing vapors.
4. double-split-ow design is shown in Fig. 3.1H. The longitudinal baes may be
solid or perforated.
5. The divided ow design (Fig. 3.1J), mechanically is like the one-pass shell ex-
cept for the addition of a nozzle. Divided ow is used to meet low-pressure-drop
requirements. The kettle reboiler is shown in Fig. 3.1K. When nucleate boiling is
to be done on the shell-side, this common design provides adequate dome space for
separation of vapor and liquid above the tube bundle and surge capacity beyond
the weir near the shell cover.
3.6 Baes and tube bundles
3.6.1 The tube bundle
Tube bundle is the most important part of a tubular heat exchanger. The tubes generally
constitute the most expensive component of the exchanger and are the one most likely to
corrode. Tube sheets, baes, or support plates, tie rods, and usually spacers complete
the bundle.
3.6.2 Bae
Baes are used to direct the side and tube side ows so that the uid velocity is increased
to obtain higher heat transfer rate and reduce fouling deposits. In horizontal units bae
are used to provide support against sagging and vibration damage. There are dierent
types of baes:
1. segemntal
2. disc and doughnut
3. orice
4. rod type
5. nest type
6. longitudinal
7. impingment
1. Segmental Baes Segmental or cross-ow baes are standard. Single, double,
and triple segmental baes are used. Bae cuts are illustrated in Fig. 3.16a. The
double segmental bae reduces crossow velocity for a given bae spacing. The
triple segmental bae reduces both cross-ow and long-ow velocities and has been
identied as the window-cut bae.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.6 Baes and tube bundles 49
a
b
c
d
Figure 3.16. Types of bae used in shell and tube heat exchanger. (a) Segmental. (b)
Segmental and strip. (c) Disc and doughnut. (d) Orice.
Minimum bae spacing is generally one-fth of the shell diameter and not less
than 50.8 mm (2 in). Maximum bae spacing is limited by the requirement to
provide adequate support for the tubes. The maximum unsupported tube span
in inches equals 74d
0.75
(where d is the outside tube diameter in inches). The
unsupported tube span is reduced by about 12 percent for aluminum, copper, and
their alloys.
Baes are provided for heat-transfer purposes. When shell-side baes are not
required for heat-transfer purposes, as may be the case in condensers or reboilers,
tube supports are installed.
Maximum bae cut is limited to about 45 percent for single segmental baes so
that every pair of baes will support each tube. Tube bundles are generally pro-
vided with baes cut so that at least one row of tubes passes through all the baes
or support plates. These tubes hold the entire bundle together. In pipe-shell ex-
changers with a horizontal bae cut and a horizontal pass rib for directing tube
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
50 3 Code and standards
side ow in the channel, the maximum bae cut, which permits a minimum of one
row of tubes to pass through all baes, is approximately 33 percent in small shells
and 40 percent in larger pipe shells.
Maximum shell-side heat-transfer rates in forced convection are apparently obtained
by cross-ow of the uid at right angles to the tubes. In order to maximize this
type of ow some heat exchangers are built with segmental-cut baes and with no
tubes in the window (or the bae cutout). Maximum bae spacing may thus equal
maximum unsupported-tube span, while conventional bae spacing is limited to
one-half of this span.
The maximum bae spacing for no tubes in the window of single segmental baes
is unlimited when intermediate supports are provided. These are cut on both sides
of the bae and therefore do not aect the ow of the shell-side uid. Each support
engages all the tubes; the supports are spaced to provide adequate support for the
tubes.
2. Rod Baes Rod or bar baes (g. 3.17) have either rods or bars extending
through the lanes between rows of tubes. A bae set can consist of a bae with
rods in all the vertical lanes and another bae with rods in all the horizontal lanes
between the tubes. The shell-side ow is uniform and parallel to the tubes. Stagnant
areas do not exist.
One device uses four baes in a bae set. Only half of either the vertical or the
horizontal tube lanes in a bae have rods. The new design apparently provides a
maximum shell-side heat-transfer coecient for a given pressure drop.
Figure 3.17. Rod baes.
3. Impingement Bae The tube bundle is customarily protected against impinge-
ment by the incoming uid at the shell inlet nozzle when the shell-side uid is at a
high velocity, is condensing, or is a twophase uid. Minimum entrance area about
the nozzle is generally equal to the inlet nozzle area. Exit nozzles also require ade-
quate area between the tubes and the nozzles. A full bundle without any provision
for shell inlet nozzle area can increase the velocity of the inlet uid by as much as
300 percent with a consequent loss in pressure.
Impingement baes are generally made of rectangular plate, although circular plates
(Fig. 3.18) are more desirable. Rods and other devices are sometimes used to
protect the tubes from impingement. In order to maintain a maximum tube count
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.6 Baes and tube bundles 51
the impingement plate is often placed in a conical nozzle opening or in a dome cap
above the shell.
Impingement baes or ow-distribution devices are recommended for axial tube-
side nozzles when entrance velocity is high.
(a)
(B)
(c)
(d)
Figure 3.18. Impingment baess;(a)Flat plate (b)curved plate (c)expanded or ared nozzle
(d) jacket type.
4. Longitudinal Flow Baes In xed-tube-sheet construction with multipass shells,
the bae is usually welded to the shell and positive assurance against bypassing
results. Removable tube bundles have a sealing device between the shell and the
longitudinal bae. Flexible light-gauge sealing strips and various packing devices
have been used. Removable U-tube bundles with four tube-side passes and two
shell-side passes can be installed in shells with the longitudinal bae welded in
place.
In split-ow shells the longitudinal bae may be installed without a positive seal
at the edges if design conditions are not seriously aected by a limited amount of
bypassing.
Fouling in petroleum-renery service has necessitated rough treatment of tube bun-
dles during cleaning operations. Many reneries avoid the use of longitudinal baes,
since the sealing devices are subject to damage during cleaning and maintenance
operations.
3.6.3 Vapor Distribution
Relatively large shell inlet nozzles, which may be used in condensers under low pressure
or vacuum, require provision for uniform vapor distribution.
3.6.4 Tube-Bundle Bypassing
Shell-side heat-transfer rates are maximized when bypassing of the tube bundle is at a
minimum. The most signicant bypass stream is generally between the outer tube limit
and the inside of the shell. The clearance between tubes and shell is at a minimum for
xed-tube-sheet construction and is greatest for straight-tube removable bundles. Ar-
rangements to reduce tube-bundle bypassing include:
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
52 3 Code and standards
1. Dummy tubes. These tubes do not pass through the tube sheets and can be
located close to the inside of the shell.
2. Tie rods with spacers. These hold the baes in place but can be located to
prevent bypassing.
3. Sealing strips. These longitudinal strips either extend from bae to bae or may
be inserted in slots cut into the baes.
4. Dummy tubes or tie rods with spacers may be located within the pass partition
lanes (and between the bae cuts) in order to ensure maximum bundle penetration
by the shell-side uid.
When tubes are omitted from the tube layout to provide entrance area about an
impingement plate, the need for sealing strips or other devices to cause proper
bundle penetration by the shell-side uid is increased.
3.6.5 Tie Rods and Spacers
Tie rods are used to hold the baes in place with spacers, which are pieces of tubing or
pipe placed on the rods to locate the baes. Occasionally baes are welded to the tie
rods, and spacers are eliminated. Properly located tie rods and spacers serve both to hold
the bundle together and to reduce bypassing of the tubes.
In very large xed-tube-sheet units, in which concentricity of shells decreases, baes are
occasionally welded to the shell to eliminate bypassing between the bae and the shell.
Metal baes are standard. Occasionally plastic baes are used either to reduce corrosion
or in vibratory service, in which metal baes may cut the tubes.
Tube plate
baffle
Spacer
Rods
Figure 3.19. Bae spacers and tie rods.
3.6.6 Tubesheets
Tubesheets are usually made from a round at piece of metal with holes drilled for the
tube ends in a precise location and pattern relative to one another. Tube sheet materials
range as tube materials. Tubes are attached to the tube sheet by pneumatic or hydraulic
pressure or by roller expansion. Tube holes can be drilled and reamed and can be machined
with one or more grooves. This greatly increases the strength of the tube joint.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
3.6 Baes and tube bundles 53
0.4mm
3 mm
a
b c
Figure 3.20. Tube sheet joint
The tubesheet is in contact with both uids and so must have corrosion resistance al-
lowances and have metalurgical and electrochemical properties appropriate for the uids
and velocities. Low carbon steel tube sheets can include a layer of a higher alloy metal
bonded to the surface to provide more eective corrosion resistance without the expense
of using the solid alloy. The tube hole pattern or pitch varies the distance from one tube
to the other and angle of the tubes relative to each other and to the direction of ow. This
allows the manipulation of uid velocities and pressure drop, and provides the maximum
amount of turbulance and tube surface contact for eective heat transfer. Where the
tube and tube sheet materials are joinable, weldable metals, the tube joint can be further
strengthened by applying a seal weld or strength weld to the joint. A strength weld has
a tube slightly reccessed inside the tube hole or slightly extended beyond the tube sheet.
The weld adds metal to the resulting lip. A seal weld is specied to help prevent the
shell and tube liquids from intermixing. In this treatment, the tube is ush with the tube
sheet surface. The weld does not add metal, but rather fuses the two materials. In cases
where it is critical to avoid uid intermixing, a double tube sheet can be provided. In this
design, the outer tube sheet is outside the shell circuit, virtually eliminating the chance
of uid intermixing. The inner tube sheet is vented to atmosphere so any uid leak is
easily detected.
Mechanisms of attaching tubes to tube sheet
Rolled Tube Joints Expanded tube-to-tube-sheet joints are standard. Properly
rolled joints have uniform tightness to minimize tube fractures, stress corrosion,
tube-sheet ligament pushover and enlargement, and dishing of the tube sheet. Tubes
are expanded into the tube sheet for a length of two tube diameters, or 50 mm (2
in), or tube-sheet thickness minus 3 mm (1/8 in). Generally tubes are rolled for the
last of these alternatives. The expanded portion should never extend beyond the
shell-side face of the tube sheet, since removing such a tube is extremely dicult.
Methods and tools for tube removal and tube rolling were discussed by John, 1959.
Tube ends may be projecting, ush, ared, or beaded (listed in order of usage). The
are or bell-mouth tube end is usually restricted to water service in condensers and
serves to reduce erosion near the tube inlet.
For moderate general process requirements at gauge pressures less than 2058 kPa
(300 lbf/in2) and less than 177
C (350
(R
2
+1)
_
2S
_
R+1
(R
2
+1)
_
_
_
_
(4.6)
where
R =
T
1
T
2
t
2
t
1
, S =
t
2
t
1
T
1
t
1
(4.7)
or in words
R =
Range of shell fluid
Range of tube fluid
, S =
Range of tube fluid
Maximum temperature difference
(4.8)
the derivation of the equation 4.6 is given by Kern (1950). The equation can be
used for any exchanger with an even number of tube passes and is plotted in Fig.4.4.
The correction factor for 2 shell passes and 4 or multiple of 4 tube passes is
F =
_
R2+1
2(R1)
_
ln
1S
1RS
ln
2/S1R+(2/S)
(1S)(1RS)+
R
2
+1
2/S1R+(2/S)
(1S)(1RS)
R
2
+1
(4.9)
These equations are plotted on g.4.4
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
58 4 Basic Design Equations of Heat Exchangers
Example 1 For example calculate the correction factor for
1. 1-2 shell and tube heat exchanger and
2. 2-4 shell and tube heat exchanger
using the equation and the graph.
T
1
= 35
o
C, T
2
= 65
o
C, t
1
= 110
o
C, t
2
= 75
o
C
R =
T
1
T
2
t
2
t
1
=
35 65
75 110
= 0.86, S =
t
2
t
1
T
1
t
1
=
75 110
35 110
= 0.467 (4.10)
From the graph of g.4.4
1. for 1-2 shell and tube heat exchanger F=0.92
2. for 2-4 shell and tube heat exchanger F=0.98
T
2
t
1
t
2
1-2 Shell and Tube
T
1
t
1
T
1
T
2
t
2
2-4 Shell and Tube
Figure 4.3. Temperature distribution for 1-2 and 2-4 shell and tube heat exchanger
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
4.1 LMTD-Method 59
Figure 4.4. Temperature correction factor: one shell, 2 shell pass, divide ow shell and split
ow shell and cross ow
4.1.3 Overall heat transfer coecient
Typical values of the overall heat transfer coecient for various types of heat exchnager
are given in . More expensive data can be found in in
The determination of U is often tedious and needs data not yet available in preliminary
stages of the design. Therefore, typical values of U are useful for quickly estimating the
required surface area. The literature has many tabulations of such typical coecients for
commercial heat transfer services.
Following is a table 4.1 with values for dierent applications and heat exchanger types.
More values can be found in the books as [29],[127], [113], [79], [93] and [14]
The ranges given in the table are an indication for the order of magnitude. Lower values
are for unfavorable conditions such as lower ow velocities, higher viscosities, and addi-
tional fouling resistances. Higher values are for more favorable conditions. Coecients
of actual equipment may be smaller or larger than the values listed. Note that the val-
ues should not be used as a replacement of rigorous methods for the nal design of heat
exchangers, although they may serve as a useful check on the results obtained by these
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
60 4 Basic Design Equations of Heat Exchangers
methods.
Table 4.1. Typical overall coecient
Hot Fluid Cold uid U (W/m
2 o
C)
Heat exchangers
Water Water 800-1500
Organic solvents organic solvent 100-300
light oils light oils 100-400
heavy oils heavy oils 50-300
Gases gass 10-50
Coolers
Organic solvents water 250-750
light oils water 350-900
heavy oils water60-900
gase water 20-300
organic solvent brine 150-500
water brine 600-1200
Gases Brine 15-250
Heaters
Steam Water 1500-4000
Steam organic solvent 500-1000
Steam light oils 300-900
Steam heavy oils 60-450
Steam gass 30-300
Dowtherm Heavy oils 50-300
Dowtherm Gases 20-200
ue gases steam 30-100
ue gases hydrocarbon vapor 30-100
Condensers
Aqueous vapor water 1000-1500
Organic vapor Water 700-1000
Organic (some non condensable gases) Water 500-700
Vacuum condensers Water 200-500
Vaporizers
Steam Aqueuos solutions 1000-1500
Steam Light organics 900-1200
Steam Heavy organics 600-900
Alternatively the overall heat transfer coecient is evalauted from the individual heat
transfer coecient as:
1
U
o
=
1
h
o
+
1
h
od
+
d
o
ln (d
o
/d
i
)
2k
w
+
d
o
d
i
1
h
i
+
d
o
d
i
1
h
id
(4.11)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
4.2 - NTU 61
where
U
o
= the overall coecient based on the outside area of the tubeW/m
2 o
C
h
o
= outside uid lm coecient, W/m
2 o
C
h
i
= inside uid lm coecient, W/m
2 o
C
h
od
= outside dirt coecient (Fouling factor), W/m
2 o
C
hi = inside dirt coecient,W/m
2 o
C
k
w
= thermal conductivity of the tube wall material, W/m
o
C
d
o
= tube outside diameter, m
d
i
= tube inside diameter, m
4.1.4 Heat transfer coecient
The heat transfer coecient is governed by general function for forced convective as
Nu =
hd
k
= f
_
Re, Pr,
d
L
,
w
_
(4.12)
and for natural convection as
Nu =
hd
k
= f
_
Gr, Pr,
w
_
(4.13)
Design equations for the heat transfer coecient for various ow geometry (tube, plate)
and conguration are given in Appendix 1. Design equation for the heat transfer coecient
for condensation and boiling is given also in appendix A.
4.1.5 Fouling factor (h
id
, h
od
)
Heat transfer may be degraded in time by corrosion, deposits of reaction products, or-
ganic growths, etc. These eects are accounted for quantitatively by fouling resistances.
Extensive data on fouling factor are given TEMA standards. Typical fouling factors for
common process and service uids are given in the table 4.2. These values are for shell
and tube heat exchangers with plain (not nned) tubes.
4.2 - NTU
The eectiveness () of a heat exchanger is dened as the ratio between the actual heat
load to the maximum possible heat load.
=
Q
Q
max
(4.14)
This is related to the heat exchanger size and capacity as
= f(NTU, C) (4.15)
Where NTU is number of transfer unit and is dened as
NTU = N =
UA
C
min
(4.16)
and C is the heat capacity ratio dened using energy equation as:
Q = M
h
C
ph
(T
hi
T
ho
) = M
c
C
pc
(T
co
T
ci
) (4.17)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
62 4 Basic Design Equations of Heat Exchangers
Table 4.2. Fouling factor
Fluid Coecient (W/m
2 o
C) Factor (resistance (m
2 o
C/W)
River water 3000-12000 0.003-0.0001
Sea water 1000-3000 0.001-0.0003
cooling water (towers) 3000-6000 0.0003-0.00017
Towns water (soft) 3000-5000 0.0003-0.0002
Towns water (hard) 1000-2000 0.001-0.0005
Steam condensate 1500-5000 0.00067-0.0002
Steam oil free 4000-10000 0.0025-0.00001
Steam oil traces 2000-5000 0.0005-0.0002
Refrigerated brine 3000-5000 0.0003-0.0002
Air and industrial gases 5000-10000 0.0002-0.00001
Flue gases 2000-5000 0.0005-0.0002
Organic vapor 5000 0.0002
Organic liquids 5000 0.0002
Light hydrocarbons 5000 0.0002
Heavy hydrocarbons 2000 0.0005
Boiling organics 2500 0.0004
Condensing organics 5000 0.0002
Heavy transfer uids 5000 0.0002
Aqueous salt solutions 3000-5000 0.0003-0.0002
M
h
C
ph
< M
c
C
pc
C
min
= M
h
C
ph
, C
max
= M
c
C
pc
(4.18)
M
h
C
pc
> M
c
C
pc
C
min
= M
c
C
pc
, C
max
= M
h
C
ph
(4.19)
Q
max
= C
min
(T
hi
T
ci
) (4.20)
C =
C
min
C
max
(4.21)
h
=
T
hi
T
ho
T
hi
T
ci
,
c
=
T
co
T
ci
T
hi
T
ci
(4.22)
=
T
c
T
span
(4.23)
where T
span
is dened in g. 4.5 for counter current ow
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
4.2 - NTU 63
Tci
Tco
Thi
Tho
T
span
0
A
Figure 4.5. Temperature distribution in counter current ow
The equation for various heat exchanger conguration is given as
Parallel ow
=
1 exp [N(1 + C)]
1 +C
(4.24)
Counter current ow
=
1 exp [N(1 + C)]
1 C exp [N(1 C)]
(4.25)
Cross ow
1. Both uid unmixed mixed
= 1 exp
_
exp(NCn) 1
Cn
_
(4.26)
where
n = N
0.22
(4.27)
2. Both uid mixed
=
_
1
1 exp(N) 1
+
C
1 exp(NC) 1
1
N
_
1
(4.28)
3. C
max
mixed, C
min
unmixed
=
1
C
{1 exp [C (1 exp(N))]} (4.29)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
64 4 Basic Design Equations of Heat Exchangers
4. C
max
unmixed, C
min
mixed
= 1 exp
_
1
C
[1 exp(NC)]
_
(4.30)
One shell pass, 2,4, 6 tube passes
= 2
_
_
1 + C +
_
(1 + C
2
)
1 + exp
_
N
_
(1 + C
2
)
_
1 exp
_
N
_
(1 +C
2
)
_
_
_
1
(4.31)
Condenser
= 1 e
N
(4.32)
Evaporator
= 1 e
N
(4.33)
Alternatively these equations are presented in a graphical form. The various curves of
vs NTU can be found in textbooks like Kern (1964( and Perry and Green (2000).
4.3 Link between LMTD and NTU
Cocurrent
ln
_
T
1
T
2
_
= ln
_
T
hi
T
ci
T
ho
T
co
_
= N
h
+ N
c
(4.34)
Counter current
ln
_
T
1
T
2
_
= ln
_
T
hi
T
co
T
ho
T
ci
_
= N
h
N
c
(4.35)
4.4 The Theta Method
Alternative method of representing the performance of heat exchangers may be given by
Theta method [146] as
=
T
m
T
span
(4.36)
where T
m
is the mean temperature dierence and T
span
is the maximum temperature
dierence (T
hi
T
ci
) (see Fig. 4.5). The Theta method is related is related to the associated
and NTU methods by expressions
=
T
m
T
span
=
NTU
(4.37)
The relationship between parameters are often presented in graphical form as shown in
Fig.4.6. However, they all depend on nding T
m
or T
lm
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
4.4 The Theta Method 65
Figure 4.6. correction charts for mean temperature dierence: (a) One shell pass and any
multiple of two tube passes. (b) Two shell passes and any multiple of four tube passes.[121].
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
66 5 Thermal Design
5 Thermal Design
5.1 Design Consideration
5.1.1 Fluid Stream Allocations
There are a number of practical guidelines which can lead to the optimum design of a
given heat exchanger. Remembering that the primary duty is to perform its thermal duty
with the lowest cost yet provide excellent in service reliability, the selection of uid stream
allocations should be of primary concern to the designer. There are many trade-os in
uid allocation in heat transfer coecients, available pressure drop, fouling tendencies
and operating pressure.
The higher pressure uid normally ows through the tube side. With their small
diameter and nominal wall thicknesses, they are easily able to accept high pressures
and avoids more expensive, larger diameter components to be designed for high
pressure. If it is necessary to put the higher pressure stream in the shell, it should
be placed in a smaller diameter and longer shell.
Place corrosive uids in the tubes, other items being equal. Corrosion is resisted
by using special alloys and it is much less expensive than using special alloy shell
materials. Other tube side materials can be clad with corrosion resistant materials
or epoxy coated.
Flow the higher fouling uids through the tubes. Tubes are easier to clean using
common mechanical methods.
Because of the wide variety of designs and congurations available for the shell
circuits, such as tube pitch, bae use and spacing, multiple nozzles, it is best to
place uids requiring low pressure drops in the shell circuit.
The uid with the lower heat transfer coecient normally goes in the shell circuit.
This allows the use of low-n tubing to oset the low transfer rate by providing
increased available surface.
Quiz: The top product of a distillation column is condensed using sea water. Allocate
the uids in the tube and the shell of the heat exchanger?.
5.1.2 Shell and tube velocity
High velocities will give high heat transfer coecients but also a high pressure drop and
cause erosion. The velocity must be high enough to prevent any suspended solids settling,
but not so high as to cause corrosion. High velocities will reduce fouling. Plastic inserts
are sometimes used to reduce erosion at the tube inlet. Typical design velocity are given
below:
Liquids
1. Tube-side process uids:1 to 2 m/s, maximum 4 m/s if required to reduce fouling:
water 1.5 to 2.5 m/s
2. Shell side: 0.3 to 1/m/s
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
5.1 Design Consideration 67
Vapors
For vapors, the velocity used will depend on the operating pressure and uid density; the
lower values in the range given below will apply to molecular weight materials
Vacuum 50 to 70 m/s
Atmospheric pressure 10 to 30 m/s
High pressure 5 to 10 m/s
5.1.3 Stream temperature
The closer the temperature approach used (the dierence between the outlet temperature
of one stream and the inlet temperature of the other stream) the larger will be the heat
transfer area required for a given duty. The optimum value will depend on the application
and can only be determined by making an economic analysis of alternative designs. As
a general guide the greater temperature dierence should be at least 20
o
C. and the
least temperature dierence 5 to 7
o
C for cooler using cooling water and 3 to 5
o
C using
refrigerated brine. The maximum temperature rise in recirculated cooling water is limited
to around 30
o
C. Care should be taken to ensure that cooling media temperatures are kept
well above the freezing point of the process materials. When heat exchange is between
process uids for heat recovery the optimum approach temperatures will normally not be
lower than 20
o
C.
5.1.4 Pressure drop
The value suggested below can be used as a general guide and will normally give designs
that are near the optimum.
Liquids
Viscosity<1 mN s/m
2
p< 35kN/m
2
Viscosity=1 to 10mN s/m
2
p= 50-70 kN/m
2
Gas and Vapors
High vacuum 0.4-0.8 kN/m
2
Medium vacuum 0.1absolute pressure
1 to 2 bar 0.5system gauge pressure
Above 10 bar 0.1system gauge pressure
When a high-pressure drop is utilized, care must be taken to ensure that the resulting
high uid velocity does not cause erosion or ow -induced tube vibration.
5.1.5 Fluid physical properties
In the correlation used to predict heat-transfer coecients, the physical properties are
usually evaluated at the mean stream temperature. This is satisfactory when the tem-
perature change is small, but can cause a signicant error when change in temperature
is large. In these circumstances , a simple and safe procedure is to evaluate the heat
transfer coecients at the stream inlet and outlet temperatures and use the lowest of the
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
68 5 Thermal Design
two value. Alternatively, the method suggested by Frank (1978) can be used; in which
Q =
A[U
2
(T
1
t
2
) U
2
(T
2
t
1
)]
ln
_
U
2
(T
1
t
2
)
U
1
(T
2
t
1
)
_
(5.1)
where U
1
, U
2
are evaluated at the end of the exchanger.
If the variation is too large for these simple methods to be used it will be necessary
to divide the temperature-enthalpy prole into sections and evaluate the heat transfer
coecients and area required for each section.
5.2 Design data
Before discussing actual thermal design, let us look at the data that must be furnished
by the process licensor before design can begin:
1. ow rates of both streams.
2. inlet and outlet temperatures of both streams.
3. operating pressure of both streams. This is required for gases, especially if the gas
density is not furnished; it is not really necessary for liquids, as their properties do
not vary with pressure.
4. allowable pressure drop for both streams. This is a very important parameter for
heat exchanger design. Generally, for liquids, a value of 0.5-0.7 kg/cm
2
is permitted
per shell. A higher pressure drop is usually warranted for viscous liquids, especially
in the tubeside. For gases, the allowed value is generally 0.05-0.2 kg/cm
2
, with 0.1
kg/cm
2
being typical.
5. fouling resistance for both streams. If this is not furnished, the designer should
adopt values specied in the TEMA standards or based on past experience.
6. physical properties of both streams. These include viscosity, thermal conductivity,
density, and specic heat, preferably at both inlet and outlet temperatures. Viscos-
ity data must be supplied at inlet and outlet temperatures, especially for liquids,
since the variation with temperature may be considerable and is irregular (neither
linear nor log-log).
7. heat duty. The duty specied should be consistent for both the shellside and the
tubeside.
8. type of heat exchanger. If not furnished, the designer can choose this based upon
the characteristics of the various types of construction described earlier. In fact, the
designer is normally in a better position than the process engineer to do this.
9. line sizes. It is desirable to match nozzle sizes with line sizes to avoid expanders
or reducers. However, sizing criteria for nozzles are usually more stringent than for
lines, especially for the shellside inlet. Consequently, nozzle sizes must sometimes be
one size (or even more in exceptional circumstances) larger than the corresponding
line sizes, especially for small lines.
10. preferred tube size. Tube size is designated as O.D., thickness, length. Some plant
owners have a preferred O.D., thickness (usually based upon inventory considera-
tions), and the available plot area will determine the maximum tube length. Many
plant owners prefer to standardize all three dimensions, again based upon inventory
considerations.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
5.3 Tubeside design 69
11. maximum shell diameter. This is based upon tube-bundle removal requirements
and is limited by crane capacities. Such limitations apply only to exchangers with
removable tube bundles, namely U-tube and oating-head. For xed-tubesheet
exchangers, the only limitation is the manufas fabrication capability and the avail-
ability of components such as dished ends and anges. Thus, oating-head heat
exchangers are often limited to a shell I.D. of 1.4-1.5 m and a tube length of 6 m
or 9 m, whereas xedtubesheet heat exchangers can have shells as large as 3 m and
tubes lengths up to 12 m or more.
12. materials of construction. If the tubes and shell are made of identical materials, all
components should be of this material. Thus, only the shell and tube materials of
construction need to be specied. However, if the shell and tubes are of dierent
metallurgy, the materials of all principal components should be specied to avoid
any ambiguity. The principal components are shell (and shell cover), tubes, channel
(and channel cover), tubesheets, and baes. Tubesheets may be lined or clad.
13. special considerations. These include cycling, upset conditions, alternative operating
scenarios, and whether operation is continuous or intermittent.
5.3 Tubeside design
Tubeside calculations are quite straightforward, since tubeside ow represents a simple
case of ow through a circular conduit. Heat-transfer coecient and pressure drop both
vary with tubeside velocity, the latter more strongly so. A good design will make the best
use of the allowable pressure drop, as this will yield the highest heat-transfer coecient.
If all the tubeside uid were to ow through all the tubes (one tube pass), it would lead
to a certain velocity. Usually, this velocity is unacceptably low and therefore has to be
increased. By incorporating pass partition plates (with appropriate gasketing) in the
channels, the tubeside uid is made to ow several times through a fraction of the total
number of tubes. Thus, in a heat exchanger with 200 tubes and two passes, the uid ows
through 100 tubes at a time, and the velocity will be twice what it would be if there were
only one pass. The number of tube passes is usually one, two, four, six, eight, and so on.
5.3.1 Heat-transfer coecient
The tubeside heat-transfer coecient is a function of the Reynolds number, the Prandtl
number, and the tube diameter. These can be broken down into the following fundamen-
tal parameters: physical properties (namely viscosity, thermal conductivity, and specic
heat); tube diameter; and, very importantly, mass velocity.
The variation in liquid viscosity is quite considerable; so, this physical property has the
most dramatic eect on heat-transfer coecient. The fundamental equation for turbulent
heat-transfer inside tubes is:
Nu = CRe
a
Pr
b
_
w
_
c
, (5.2)
or
h = C
k
D
_
GD
_
a _
C
p
k
_
b
_
w
_
c
(5.3)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
70 5 Thermal Design
where
Nu =
hd
e
k
Nusselt number
Pr =
Cp
k
Prandtl number
Re
ud
Reynolds number
d
e
4A
P
hydraulic diameter
A cross-sectional area
P wetted perimeter
u uid velocity
w
uid viscosity at the tube wall temperature
k uid thermal conductivity
C
p
uid specic heat
C =
_
_
0.021 gases
0.023 non-viscous liquid
0.027 viscous liquid
a = 0.8
b = 0.3 for cooling
b = 0.4 for heating
c = 0.14
Viscosity inuences the heat-transfer coecient in two opposing ways- as a parameter of
the Reynolds number, and as a parameter of Prandtl number. Thus, from Eq. 5.3:
h
0.80.33
=
0.47
(5.4)
In other words, the heat-transfer coecient is inversely proportional to viscosity to the
0.47 power. Similarly, the heat-transfer coecient is directly proportional to thermal
conductivity to the 0.67 power.
These two facts lead to some interesting generalities about heat transfer. A high thermal
conductivity promotes a high heat-transfer coecient. Thus, cooling water (thermal
conductivity of around 0.55 kcal/hm
C) at 250-1,300 kcal/hm
2
C, and then hydrocarbon gases
(thermal conductivity between 0.02 and 0.03 kcal/hm
C) at 50-500 kcal/hm
2
C.
Hydrogen is an unusual gas, because it has an exceptionally high thermal conductivity
(greater than that of hydrocarbon liquids). Thus, its heat-transfer coecient is toward
the upper limit of the range for hydrocarbon liquids.
The range of heat-transfer coecients for hydrocarbon liquids is rather large due to the
large variation in their viscosity, from less than 0.1 cP for ethylene and propylene to more
than 1,000 cP or more for bitumen. The large variation in the heat-transfer coecients
of hydrocarbon gases is attributable to the large variation in operating pressure. As
operating pressure rises, gas density increases. Pressure drop is directly proportional to
the square of mass velocity and inversely proportional to density. Therefore, for the same
pressure drop, a higher mass velocity can be maintained when the density is higher. This
larger mass velocity translates into a higher heat-transfer coecient.
5.3.2 Pressure drop
The pressure drop due to friction exists because of the shear stress between the uid and
the tube wall. Estimation of the friction pressure drop is somewhat more complex and
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
5.3 Tubeside design 71
various approaches have been taken, for example the frictional pressure gradient is given
as
_
dp
dz
_
f
=
4
o
d
=
4fG
2
2d
, (5.5)
where G is the mass ux in kg/m
2
s and f is the friction factor calculated using a Blasius-
type model as
f =
_
_
0.3164
Re
0.25
Re 2320
64
Re
Re < 2320 .
Integration of equation B.1 yields
p =
4fG
2
2
L
d
, (5.6)
Mass velocity strongly inuences the heat-transfer coecient. For turbulent ow, the
tubeside heat-transfer coecient varies to the 0.8 power of tubeside mass velocity, whereas
tubeside pressure drop varies to the square of mass velocity. Thus, with increasing mass
velocity, pressure drop increases more rapidly than does the heat-transfer coecient.
Consequently, there will be an optimum mass velocity above which it will be wasteful to
increase mass velocity further.
Furthermore, very high velocities lead to erosion. However, the pressure drop limitation
usually becomes controlling long before erosive velocities are attained. The minimum
recommended liquid velocity inside tubes is 1.0 m/s, while the maximum is 2.5-3.0 m/s.
Pressure drop is proportional to the square of velocity and the total length of travel.
Thus, when the number of tube passes is increased for a given number of tubes and a
given tubeside ow rate, the pressure drop rises to the cube of this increase. In actual
practice, the rise is somewhat less because of lower friction factors at higher Reynolds
numbers, so the exponent should be approximately 2.8 instead of 3.
Tubeside pressure drop rises steeply with an increase in the number of tube passes. Con-
sequently, it often happens that for a given number of tubes and two passes, the pressure
drop is much lower than the allowable value, but with four passes it exceeds the allowable
pressure drop. If in such circumstances a standard tube has to be employed, the designer
may be forced to accept a rather low velocity. However, if the tube diameter and length
may be varied, the allowable pressure drop can be better utilized and a higher tubeside
velocity realized.
The following tube diameters are usually used in the CPI: (1/4, 3/8, 1/2, 5/8, 3/4, 1, 1
1/4, 1 1/2 in. Of these, 3/4 in. and 1 in. are the most popular. Tubes smaller than 3/4
in. O.D. should not be used for fouling services. The use of small-diameter tubes, such as
1 in., is warranted only for small heat exchangers with heat-transfer areas less than 20-30
m
2
.
It is important to realize that the total pressure drop for a given stream must be met.
The distribution of pressure drop in the various heat exchangers for a given stream in a
particular circuit may be varied to obtain good heat transfer in all the heat exchangers.
Consider a hot liquid stream owing through several preheat exchangers. Normally, a
pressure drop of 0.7 kg/cm
2
per shell is permitted for liquid streams. If there are ve
such preheat exchangers, a total pressure drop of 3.5 kg/cm
2
for the circuit would be
permitted. If the pressure drop through two of these exchangers turns out to be only 0.8
kg/cm
2
, the balance of 2.7 kg/cm
2
would be available for the other three.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
72 5 Thermal Design
5.4 Shell side design
Shell side design The shellside calculations are far more complex than those for the tube-
side. This is mainly because on the shellside there is not just one ow stream but one
principal cross-ow stream and four leakage or bypass streams. There are various shell-
side ow arrangements, as well as various tube layout patterns and baing designs, which
together determine the shellside stream analysis.
5.4.1 Shell conguration
TEMA denes various shell patterns based on the ow of the shellside uid through the
shell: E, F, G, H, J, K, and X (see Figure 3.1).
In a TEMA E single-pass shell, the shellside uid enters the shell at one end and leaves
from the other end. This is the most common shell type - more heat exchangers are built
to this conguration than all other con- gurations combined.
A TEMA F two-pass shell has a longitudinal bae that divides the shell into two passes.
The shellside uid enters at one end, traverses the entire length of the exchanger through
one-half the shell cross-sectional area, turns around and ows through the second pass,
then nally leaves at the end of the second pass. The longitudinal bae stops well short
of the tubesheet, so that the uid can ow into the second pass.
The F shell is used for temperature- cross situations - that is, where the cold stream leaves
at a temperature higher than the outlet temperature of the hot stream. If a two-pass (F)
shell has only two tube passes, this becomes a true countercurrent arrangement where a
large temperature cross can be achieved.
A TEMA J shell is a divided-ow shell wherein the shellside uid enters the shell at the
center and divides into two halves, one owing to the left and the other to the right and
leaving separately. They are then combined into a single stream. This is identied as a
J 1-2 shell. Alternatively, the stream may be split into two halves that enter the shell at
the two ends, ow toward the center, and leave as a single stream, which is identied as
a J 2-1 shell.
A TEMA G shell is a split-ow shell (see Figure 3.1). This construction is usually em-
ployed for horizontal thermosyphon reboilers. There is only a central support plate and
no baes. A G shell cannot be used for heat exchangers with tube lengths greater than
3 m, since this would exceed the limit on maximum unsupported tube length specied by
TEMA - typically 1.5 m, though it varies with tube O.D., thickness, and material.
When a larger tube length is needed, a TEMA H shell (see Figure3.1) is used. An H shell
is basically two G shells placed side-by-side, so that there are two full support plates. This
is described as a double-split conguration, as the ow is split twice and recombined twice.
This construction, too, is invariably employed for horizontal thermosyphon reboilers. The
advantage of G and H shells is that the pressure drop is drastically less and there are no
cross baes.
A TEMA X shell (see Figure 3.1) is a pure cross-ow shell where the shellside uid enters
at the top (or bottom) of the shell, ows across the tubes, and exits from the opposite side
of the shell. The ow may be introduced through multiple nozzles located strategically
along the length of the shell in order to achieve a better distribution. The pressure drop
will be extremely low - in fact, there is hardly any pressure drop in the shell, and what
pressure drop there is, is virtually all in the nozzles. Thus, this conguration is employed
for cooling or condensing vapors at low pressure, particularly vacuum. Full support plates
can be located if needed for structural integrity; they do not interfere with the shellside
ow because they are parallel to the ow direction.
A TEMA K shell (see Figure 3.1) is a special cross-ow shell employed for kettle reboilers
(thus the K). It has an integral vapor-disengagement space embodied in an enlarged shell.
Here, too, full support plates can be employed as required.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
5.4 Shell side design 73
5.4.2 Tube layout patterns
There are four tube layout patterns, as shown in Figure 5.1: triangular (30
), rotated
triangular (60
), square (90
).
Figure 5.1. Tubes layout pattern.
A triangular (or rotated triangular) pattern will accommodate more tubes than a square
(or rotated square) pattern. Furthermore, a triangular pattern produces high turbulence
and therefore a high heat-transfer coecient. However, at the typical tube pitch of 1.25
times the tube O.D., it does not permit mechanical cleaning of tubes, since access lanes
are not available. Consequently, a triangular layout is limited to clean shellside services.
For services that require mechanical cleaning on the shellside, square patterns must be
used. Chemical cleaning does not require access lanes, so a triangular layout may be used
for dirty shellside services provided chemical cleaning is suitable and eective.
A rotated triangular pattern seldom oers any advantages over a triangular pattern, and
its use is consequently not very popular.
For dirty shellside services, a square layout is typically employed. However, since this is an
in-line pattern, it produces lower turbulence. Thus, when the shellside Reynolds number
is low (< 2,000), it is usually advantageous to employ a rotated square pattern because
this produces much higher turbulence, which results in a higher eciency of conversion
of pressure drop to heat transfer.
As noted earlier, xed-tubesheet construction is usually employed for clean services on
the shellside, Utube construction for clean services on the tubeside, and oating-head
construction for dirty services on both the shellside and tubeside. (For clean services
on both shellside and tubeside, either xed-tubesheet or U-tube construction may be
used, although U-tube is preferable since it permits dierential expansion between the
shell and the tubes.) Hence, a triangular tube pattern may be used for xed-tubesheet
exchangers and a square (or rotated square) pattern for oating-head exchangers. For
U-tube exchangers, a triangular pattern may be used provided the shellside stream is
clean and a square (or rotated square) pattern if it is dirty.
5.4.3 Tube pitch
Tube pitch is dened as the shortest distance between two adjacent tubes.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
74 5 Thermal Design
For a triangular pattern, TEMA species a minimum tube pitch of 1.25 times the tube
O.D. Thus, a 25- mm tube pitch is usually employed for 20-mm O.D. tubes.
For square patterns, TEMA additionally recommends a minimum cleaning lane of 4 in.
(or 6 mm) between adjacent tubes. Thus, the minimum tube pitch for square patterns
is either 1.25 times the tube O.D. or the tube O.D. plus 6 mm, whichever is larger. For
example, 20-mm tubes should be laid on a 26-mm (20 mm + 6 mm) square pitch, but
25-mm tubes should be laid on a 31.25-mm (25 mm 1.25) square pitch.
Designers prefer to employ the minimum recommended tube pitch, because it leads to
the smallest shell diameter for a given number of tubes. However, in exceptional cir-
cumstances, the tube pitch may be increased to a higher value, for example, to reduce
shellside pressure drop. This is particularly true in the case of a cross-ow shell.
5.4.4 Baing
Type of baes. Baes are used to support tubes, enable a desirable velocity to be
maintained for the shellside uid, and prevent failure of tubes due to ow-induced vibra-
tion. There are two types of baes: plate and rod. Plate baes may be single-segmental,
double-segmental, or triple-segmental, as shown in Figure 5.2.
Figure 5.2. Types of baes.
Bae spacing. Bae spacing is the centerline-to-centerline distance between adjacent
baes. It is the most vital parameter in STHE design.
The TEMA standards specify the minimum bae spacing as one-fth of the shell inside
diameter or 2 in., whichever is greater. Closer spacing will result in poor bundle pene-
tration by the shellside uid and diculty in mechanically cleaning the outsides of the
tubes. Furthermore, a low bae spacing results in a poor stream distribution as will be
explained later.
The maximum bae spacing is the shell inside diameter. Higher baf- e spacing will
lead to predominantly longitudinal ow, which is less ecient than cross-ow, and large
unsupported tube spans, which will make the exchanger prone to tube failure due to
ow-induced vibration.
Optimum bae spacing. For turbulent ow on the shellside (Re > 1,000), the heat-
transfer coecient varies to the 0.6-0.7 power of velocity; however, pressure drop varies
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
5.4 Shell side design 75
to the 1.7-2.0 power. For laminar ow (Re < 100), the exponents are 0.33 for the heat-
transfer coecient and 1.0 for pressure drop. Thus, as bae spacing is reduced, pressure
drop increases at a much faster rate than does the heat-transfer coecient.
This means that there will be an optimum ratio of bae spacing to shell inside diameter
that will result in the highest eciency of conversion of pressure drop to heat transfer.
This optimum ratio is normally between 0.3 and 0.6.
Bae cut. As shown in Figure 5.3, bae cut is the height of the segment that is cut in
each bae to permit the shellside uid to ow across the bae. This is expressed as a
percentage of the shell inside diameter. Although this, too, is an important parameter
for STHE design, its eect is less profound than that of bae spacing.
Figure 5.3. Bae cut.
Bae cut can vary between 15% and 45% of the shell inside diameter.
Both very small and very large bae cuts are detrimental to ecient heat transfer on the
shellside due to large deviation from an ideal situation, as illustrated in Figure 5.4.
Figure 5.4. Eect of small and large bae cuts.
It is strongly recommended that only bae cuts between 20% and 35% be employed. Re-
ducing bae cut below 20% to increase the shellside heat-transfer coecient or increasing
the bae cut beyond 35% to decrease the shellside pressure drop usually lead to poor de-
signs. Other aspects of tube bundle geometry should be changed instead to achieve those
goals. For example, doublesegmental baes or a divided-ow shell, or even a cross-ow
shell, may be used to reduce the shellside pressure drop.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
76 5 Thermal Design
For single-phase uids on the shellside, a horizontal bae cut (Figure 5.5) is recommended,
because this minimizes accumulation of deposits at the bottom of the shell and also
prevents stratication. However, in the case of a two-pass shell (TEMA F), a vertical cut
is preferred for ease of fabrication and bundle assembly.
Figure 5.5. Bae cut orientation
5.4.5 Equalize cross-ow and window velocities
Flow across tubes is referred to as cross-ow, whereas ow through the window area (that
is, through the bae cut area) is referred to as window ow.
The window velocity and the cross-ow velocity should be as close as possible - preferably
within 20%
of each other. If they dier by more than that, repeated acceleration and deceleration take
place along the length of the tube bundle, resulting in inecient conversion of pressure
drop to heat transfer.
5.4.6 Shellside stream analysis (Flow pattern)
On the shellside, there is not just one stream, but a main cross-ow stream and four
leakage or bypass streams, as illustrated in Figure 5.6. Tinker (4) proposed calling these
streams the main cross-ow stream (B), a tube-to-bae-hole leakage stream (A), a bundle
bypass stream (C), a pass-partition bypass stream (F), and a bae-to-shell leakage stream
(E). While the B (main cross-ow) stream is highly eective for heat transfer, the other
streams are not as eective. The A stream is fairly ecient, because the shellside uid
is in contact with the tubes. Similarly, the C stream is in contact with the peripheral
tubes around the bundle, and the F stream is in contact with the tubes along the pass-
partition lanes. Consequently, these streams also experience heat transfer, although at
a lower eciency than the B stream. However, since the E stream ows along the shell
wall, where there are no tubes, it encounters no heat transfer at all.
The fractions of the total ow represented by these ve streams can be determined for a
particular set of exchanger geometry and shellside ow conditions by any sophisticated
heatexchanger thermal design software. Essentially, the ve streams are in parallel and
ow along paths of varying hydraulic resistances. Thus, the ow fractions will be such that
the pressure drop of each stream is identical, since all the streams begin and end at the
inlet and outlet nozzles. Subsequently, based upon the eciency of each of these streams,
the overall shellside stream eciency and thus the shellside heat-transfer coecient is
established.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
5.4 Shell side design 77
Figure 5.6. Tube arrangement
Since the ow fractions depend strongly upon the path resistances, varying any of the
following construction parameters will aect stream analysis and thereby the shellside
performance of an exchanger:
bae spacing and bae cut;
tube layout angle and tube pitch;
number of lanes in the ow direction and lane width;
clearance between the tube and the bae hole;
clearance between the shell I.D. and the bae; and
location of sealing strips and sealing rods.
Using a very low bae spacing tends to increase the leakage and bypass streams. This
is because all ve shellside streams are in parallel and, therefore, have the same pressure
drop. The leakage path dimensions are xed. Consequently, when bae spacing is de-
creased, the resistance of the main cross-ow path and thereby its pressure drop increases.
Since the pressure drops of all ve streams must be equal, the leakage and bypass streams
increase until the pressure drops of all the streams balance out. The net result is a rise
in the pressure drop without a corresponding increase in the heat-transfer coecient.
The shellside uid viscosity also aects stream analysis profoundly. In addition to inu-
encing the shellside heat transfer and pressure drop performance, the stream analysis also
aects the mean temperature dierence (MTD) of the exchanger. This will be discussed
in detail later. First, though, lets look at an example that demonstrates how to optimize
bae design when there is no signicant temperature prole distortion.
5.4.7 Heat transfer coecient and pressure drop
For the shell side heat transfer coecient and pressure drop there are a number of methods
these include:
Kerns method
Donohues method
Bell-Delaware method
Tinkers method
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
78 5 Thermal Design
Besides these methods there is some proprietary methods putout by various organization
for use by their member companies. A number of these method are based on one of the
above methods. Some are based upon a judicious combination of methods 3 and 4 above
and supplemented by further research data. Among the most popular of the proprietary
methods, judged by their large clientele are
Heat Transfer Research Inc. (HTRI), Alliambra, california. This method is also
known as stream analysis method.
Heat Transfer and Fluid Flow Service (HTFS), Engineering Science Division, AERE,
Harwell, United Kingdom Method.
In this work only Kerns method is given below. Bell-Delaware method may be found in
Coulson and Richardsons
5.4.8 Heat transfer coecient
Nu = 0.36Re
0.55
Pr
1/3
_
w
_
0.14
, (5.7)
where
Nu =
hde
k
Nusselt number
Pr =
C
p
k
Prandtl number
Re =
Gd
e
Reynolds number
d
e
=
4A
P
hydraulic diameter
A = cross-sectional ow area
P = wetted perimeter
G =
M
As
Mass ux
A
s
=
(ptdo)Dsl
B
pt
uid viscosity at the tube wall temperature
p
t
= pitch diameter
D
s
= shell diameter
l
B
= Bae spacing
Hydraulic diameter (Fig. 5.1)
d
e
=
_
_
p
2
t
d
2
o
/4
do
for square pitch
0.87p
2
t
/2d
2
o
/8
d
o
/2
for equilateral triangular pitch
5.4.9 Pressure drop
p = 4f
_
D
s
d
_
_
u
2
2
_
_
L
l
b
_
_
w
_
0.14
, (5.8)
where
f =
_
_
0.3164
Re
0.25
Re 2320
64
Re
Re < 2320 .
L=tube length
l
B
= bae spacing. The term (L/l
B
) is the number of times the ow crosses the tube
bundle=(N
B
+ 1). Where N
B
is the number of baes.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
5.5 Design Algorithm 79
5.5 Design Algorithm
Step1
Specification
Define duty Q
Make energy balance if needed
to calcualted unspecified flow
rates or temperature
Q=M )
c pc c2 c1 h ph h1 h2
c (T -T )=M C (T -T
Step2
Calculate physical properties
Step3
Assume value of overall
coefficient U
o,ass
Step 4
Decide number of shell and
tube passes
Calculate T T , F and
lm m
Step 5
Determine heat transfer area
required A T
o o,ass
=q/U
m
Step 6
Decide type, tube size, material,
layout
Assign fluids to shell or tube
Step 7
Calculate number of tubes
Step 8
Calculate shell diameter
Step 9
Estimate tube-side heat
transfer coefficient
Step 10
Decide baffle spacing and estimate
shell side heat transfer coefficient
Step 11
Calculate overall heat transfer
Coefficient including fouling factors
U
o,cal
Step 12
Estimate tube and shell side
pressure drop
Step 13
Estimate cost of heat exchanger
Can design be
optmized to
reduce cost?
Accept design
Is pressure drops
within specification?
0<(U
o,cal o,ass o,ass
-U )/U <30 Set U
o,ass
=U
o,cal
Yes
No
No
Figure 5.7. Design procedure for shell and tube heat exchanger.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
80 6 Specication sheet
6 Specication sheet
Specication sheet is a data sheet that contains the information provided by the customer
to the vendor for the process and mechanical designs of an exchanger. After the process
design is done, the engineer lls in some further information. The rest of the information
is lled after the mechanical design is completed. The specication sheet is a medium of
communication between dierent parties involved in the procurement, design and fabri-
cation of heat exchanger. It is also used to compare the performance of the installed unit
with the design conditions.
6.1 Information included
The information contained in the sheet is best decribed by a data sheet. Although each
company has its own version of data sheet, the most popular one is that of the TEMA
standards. It is similar to that of API standard 660. It contains the uid
ow rate and properties,
heat duty,
heat transfer coecient,
fouling resistance,
details about the shell and tube size,
materials,
bae nozzle, etc..
Some variations include information for alternate designs and dierent systems of units
(British, SI, metric).
6.2 Information not included
The regarding the type of anges, studs, vent and relief valves, drains lines, welding,
inspection and testing requirement of the material of construction, etc.. are not given in
the specication sheet.
6.3 Operation conditions
The following operating conditions regarding the exchanger operation should be known
to the thermal designer for critical application.
1. Start-up condition and procedure
2. Normal operating conditions
3. Upset and emergency conditions
4. shut down conditions and procedure
5. possibility of switching the shell-side and tube tube side uid for better design
6. possibility of increasing the allowable pressure drop to control the fouling
7. beside these the spec-sheet should provided with other information concerning the
composition of the streams, their thermal and physical properties and any phase
change occurring.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
6.4 Bid evaluation 81
6.4 Bid evaluation
6.4.1 Factor to be consider
For ease in evaluations of the bids submitted by competitive bidders, all pertinent data
from each bid should be put on a large data sheet. During evaluation the following factor
should be kept in mid:
1. The design submitted by the bidders should meet the heat transfer and pressure
drop requirements. Set the upper and lower limit of pressure drop for each bid.
2. if the designs oered by bidder vary, the spec-sheet provided to them should be
checked to see if any anomalies exist
3. Adequate vent, drainage and safety valve should be provided
4. Units should not have hot spot or dead zones
5. Information about vibration analysis must be checked
6. for fouling on the shell side, the tube lay out should permit easy cleaning
7. The fabrication shop should have a good reputation and certicate of inspection
8. The material of construction should be available at the country of the bidder or
their import should not pose any diculty
9. the delivery should be on schedule
10. cost should be low, cost escalation should be included
11. the payment, penalty, and guarantee clauses in the contact should be evenly balance
and be unduly favorable to the bidder
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
82 6 Specication sheet
Figure 6.1. Data sheet
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
83
7 Storage, Installation, Operation and Maintenance
Proper storage, installation handling and correct start up emergency, and shutdown pro-
cedure are important for the successful working of a well designed and fabricated heat
exchanger. regular cleaning, maintenance and repairs are necessary to ensure trouble free
operation of the unit for its designed life span. These will be discussed in the following
sections.
NOTE: Before placing your equipment in operation, environment and service conditions
should be checked for compatibility with materials of construction. Contact your nearest
heat exchanger Standard representative if you are not sure what the actual materials of
construction are.
Successful performance of heat transfer equipment, length of service and freedom from
operating diculties are largely dependent upon:
1. Proper thermal design.
2. Proper physical design.
3. Storage practice prior to installation.
4. Manner of installation, including design of foundation and piping.
5. The method of operation.
6. The thoroughness and frequency of cleaning.
7. The materials, workmanship, and tools used in maintenance and making repairs
and replacements.
Failure to perform properly may be due to one or more of the following:
1. Exchanger being dirty.
2. Failure to remove preservation materials after storage.
3. Operating conditions being dierent than design conditions.
4. Air or gas binding.
5. Incorrect piping connections.
6. Excessive clearances between internal parts due to corrosion.
7. Improper application.
7.1 Storage
Standard heat exchangers are protected against the elements during shipment. If they
cannot be installed and put into operation immediately upon receipt at the jobsite, cer-
tain precautions are necessary to prevent deterioration during storage. Responsibility for
integrity of the heat exchangers must be assumed by the user. The manufacturer will not
be responsible for damage, corrosion or other deterioration of heat exchanger equipment
during transit and storage.
Good storage practices are important, considering the high costs of repair or replacement,
and the possible delays for items which require long lead times for manufacture. The
following suggested practices are provided solely as a convenience to the user, who shall
make his own decision on whether to use all or any of them.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
84 7 Storage, Installation, Operation and Maintenance
1. On receipt of the heat exchanger, inspect for shipping damage to all protective cov-
ers. If damage is evident, inspect for possible contamination and replace protective
covers as required. If damage is extensive, notify the carrier immediately.
2. If the heat exchanger is not to be placed in immediate service, take precautions to
prevent rusting or contamination.
3. Heat exchangers for oil service, made of ferrous materials, may be pressure-tested
with oil at the factory. However, the residual oil coating on the inside surfaces of
the exchanger does not preclude the possibility of rust formation. Upon receipt,
ll these exchangers with appropriate oil or coat them with a corrosion prevention
compound for storage. These heat exchangers have a large warning decal, indicating
that they should be protected with oil.
4. The choice of preservation of interior surfaces during storage for other service appli-
cations depends upon your system requirements and economics. Only when included
in the original purchase order specications will specic preservation be incorporated
prior to shipment from the factory.
5. Remove any accumulations of dirt, water, ice or snow and wipe dry before moving
exchangers into indoor storage. If unit was not lled with oil or other preservative,
open drain plugs to remove any accumulated moisture, then reseal. Accumulation
of moisture usually indicates rusting has already started and remedial action should
be taken.
6. Store under cover in a heated area, if possible. The ideal storage environment for
heat exchangers and accessories is indoors, above grade, in a dry, low humidity at-
mosphere which is sealed to prevent entry of blowing dust, rain or snow. Maintain
temperatures between 70
F and 105
w
_
c
, (A.1)
where
Nu =
hd
e
k
Nusselt number
Pr =
C
p
k
Prandtl number
Re
ud
Reynolds number
d
e
4A
P
hydraulic diameter
A cross-sectional area
P wetted perimeter
u uid velocity
w
uid viscosity at the tube wall temperature
k uid thermal conductivity
C
p
uid specic heat
C =
_
_
0.021 gases
0.023 non-viscous liquid
0.027 viscous liquid
a = 0.8
b = 0.3 for cooling
b = 0.4 for heating
c = 0.14
A.1.2 Inside tube: Laminar ow
Nu = 1.86
_
RePr
d
L
_
1/3
_
w
_
0.14
, (A.2)
A.1.3 Shell side
For the shell side heat transfer coecient there are a number of methods the include:
Kerns method
Donohues method
Bell-Delaware method
Tinkers method
Besides these methods there is some proprietary methods putout by various organization
for use by their member companies. A number of these method are based on one of the
above methods. Some are based upon a judicious combination of methods 3 and 4 above
and supplemented by further research data. Among the most popular of the proprietary
methods, judged by their large clientele are
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
132 A Heat transfer coecient
Heat Transfer Research Inc. (HTRI), Alliambra, california. This method is also
known as stream analysis method.
Heat Transfer and Fluid Flow Service (HTFS), Engineering Science Division, AERE,
Harwell, United Kingdom Method.
In this work only Kerns method is given below. Bell-Delaware method may be found in
Coulson and Richardsons
Nu = 0.36Re
0.55
Pr
1/3
_
w
_
0.14
, (A.3)
where
Nu =
hde
k
Nusselt number
Pr =
C
p
k
Prandtl number
Re =
Gd
e
Reynolds number
d
e
=
4A
P
hydraulic diameter
A = cross-sectional ow area
P = wetted perimeter
G =
M
As
Mass ux
A
s
=
(ptdo)Dsl
B
pt
uid viscosity at the tube wall temperature
p
t
= pitch diameter
D
s
= shell diameter
l
B
= Bae spacing
Hydraulic diameter (Fig. A.1)
d
e
=
_
_
p
2
t
d
2
o
/4
do
for square pitch
0.87p
2
t
/2d
2
o
/8
d
o
/2
for equilateral triangular pitch
p
t
d
o
Square pitch
pt
Equilateral triangular pitch
A
s
Cross-flow area
Figure A.1. Tube arrangement
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
A.2 Condensation 133
A.1.4 Plate heat exchanger
Nu = 0.26Re
0.65
Pr
0.4
_
w
_
0.14
, (A.4)
where
Nu =
hde
k
= Nusselt number
Pr =
C
p
k
= Prandtl number
Re =
u
p
d
e
=
Gde
= Reynolds number
d
e
= hydraulic diameter, taken as twice the gap between the plates
A = cross-sectional ow area
P = wetted perimeter
G =
M
A
f
= Mass ux
A
f
= cross-sectional area for ow
u
p
= channel velocity
M = mass ow rate
A.2 Condensation
A.2.1 Condensation on vertical plate or outside vertical tube
h
m
= 0.943
_
k
3
g
TL
_
1/4
, (A.5)
where
h
m
= mean heat transfer coecient
L = lenth of the plate or the vetical tube
k thermal conductivity of the saturated liquid lm
= liquid density
= liquid viscosity
= latent heat of evaporization
T = T
s
T
w
temperature dierence across the condensate lm
g = acceleration due to gravity
T
s
= saturation temperature of the condensate lm
T
w
= wall temperature
A.2.2 Condensation on external horizontal tube
h
m
= 0.725
_
k
3
g
Td
o
_
1/4
, (A.6)
where
d
o
= out side diamter of the tube
A.2.3 Condensation on banks of horizontal tube
h
m
= 0.725
_
k
3
g
TJd
o
_
1/4
, (A.7)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
134 A Heat transfer coecient
where
J = number of tubes in a row (Fig. ??)
In the above equation the condensate lm properties save the latent heat of vaporization
are evaluated at the lm temperature.
T
f
=
T
s
+T
w
2
, (A.8)
the latent heat of vaporization is evaluated at the condensate temperature. For the case
of subcooling or superheating the heat transfer coecient is corrected by substituting the
corrected latent heat the heat transfer equation (Rohsenow et al. [121] and Carey [18]) in
Nusselt [109])
= + 0.68c
p
T . (A.9)
A.2.4 Condensation inside horizontal tube
h
m
= 0.555
_
k
3
g
Td
_
1/4
, (A.10)
A.3 Two phase ow: Pure uid
A.3.1 Steiner [140] correlation
Steiner [140] has considered the two phase heat transfer coecient h as a combination of
the convective and the nucleate part using an asymptotic model as:
h =
_
h
3
n
+ h
3
c
_
1/3
, (A.11)
where h
n
and h
c
is the nucleate and convective boiling heat transfer coecient respectively.
The convective boiling heat transfer coecient for a completely wetted tube (i.e. all types
of ow patterns save stratied and stratied-wavy ow) is calculated as
h
c
h
L0
=
_
_
_
_
_
(1 x) + 1.2 x
0.4
(1 x)
0.01
_
G
_
0.37
_
_
+
_
_
h
G0
h
L0
x
0.01
_
_
1 + 8(1 x)
0.7
_
G
_
0.67
_
_
_
_
2
_
_
0.5
. (A.12)
The heat transfer coecients h
L0
and h
G0
are those of single phase ow, assuming that
the total mass velocity is pure liquid or pure vapor respectively. They are calculated in
the case of a fully developed turbulent ow from the Gnielinski [46] model
Nu =
(/8)(Re 1000)Pr
1 + 12.7(/8)
0.5
(Pr
2/3
1)
, (A.13)
taken in to account the respective dimensionless group Nu
L0
, Nu
G0
, Re
L0
, Re
G0
, Pr
l
and
Pr
g
. These dimensionless groups are dened as
Nu
L0/G0
=
h
L0/G0
d
k
L/G
, (A.14)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
A.3 Two phase ow: Pure uid 135
Re
L0/G0
=
md
L/G
, (A.15)
Pr
L/G
=
L/G
c
p,L/G
k
L/G
. (A.16)
The friction factor is
= (1.82logRe 1.62)
2
. (A.17)
For a partial wetting of the tube (stratied or stratied-wavy ow) the average heat
transfer coecient at the tube circumference under the thermal boundary condition of a
constant wall temperature is given as
h
c
= h
wet
(1 ) + h
G
, (A.18)
where h
wet
is the convective boiling heat transfer coecient at the wetted part of the
tube and it is calculated by using equation A.12. In the non-wetted part of the tube,
the convective heat transfer coecient h
g
is calculated from the Gnielinski [46] model
(equation A.13). In this case Re and Nu are dened with the hydraulic diameter of the
vapor-occupied part of the tube cross-section
d
h
= d
_
sin
d + 2 sin(/2)
_
, (A.19)
where is the stratied angle. The Reynolds number is given as
Re
G
=
m xd
hyd
G
, (A.20)
and
h
G
=
Nu
G
k
G
d
hyd
. (A.21)
The void fraction is calculated using the Rauhani [117] model given as
=
x
G
_
(1 + 0.12(1 x))
_
x
G
+
1 x
L
_
+
1.18(1 x)[g(
L
G
)]
1/4
m .
1/2
L
_
1
(A.22)
The wetting boundary can be estimated (see Fig. A.2) from the void fraction as
=
f
G
f
G
+ f
L
. (A.23)
With some mathematical manipulation of equation A.23 the non-wetted perimeter can
calculated iteratively from the following relationship
= 2 + sin , (A.24)
with the assumption that no bubbles in the liquid phase and no entrainment (hold-up) in
the vapor phase, the scaling parameter of equation A.18 can thus be calculated as
=
G
2
, (A.25)
where
G
= 0.5.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
136 A Heat transfer coecient
d
h
f
L
f
G
U
i
U
G
U
L
Figure A.2. Cross-section and perimeter parts of the vapor ow in a horizontal tube.
The local nucleate boiling heat transfer coecient h
nb
of a horizontal tube is estimated
as
h
nb
h
o
= C
f
_
q
q
o
_
n(pr)
F(p
r
)F(R
a
)F(d)F( m, x) . (A.26)
The value with a subscript o is a reference value.
The pressure function is given as
F(p
r
) = 2.692p
0.43
r
+
_
1.6p
6.5
r
1 p
4.4
r
_
, (A.27)
and the mass ux function is given as
F( m, x) =
m
m
o
0.25
_
_
1 p
0.1
r
_
q
q
cr,nb
_
0.3
x
_
_
, (A.28)
where
q
cr,cb
= 2.79 q
cr,0,1
p
0.4
r
(1 p
r
) . (A.29)
The critical value of q
cr,0,1
at a reduced pressure p
r
of 0.1 is given as
q
cr,0.1
= 0.13h
V,0
0.5
G,0
[
o
g(
L,0
G,0
)]
0.25
. (A.30)
The function for the eect of surface roughness and tube diameter is F(R
a
) =(R
a
/R
ao
)
0.133
and F(d)=(d
o
/d)
0.5
respectively. The pressure dependence of the heat ux exponent n(p
r
)
can be predicted as
n(p
r
) = 0.9 0.3p
0.3
r
. (A.31)
The experimental value of the specic constant C
f
for a number of substances is be found
in VDI-Warmeatlas[157], for example for water it is 0.72. In absence of an experimental
value it can be estimated as
C
f
= 0.789
_
M
M
H
2
_
0.11
, (A.32)
where
M is the molecular weight and
M
H
2
= 2.016. The correction factor for a stratied
and a stratied-wavy ow pattern under the thermal boundary condition of a constant
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
A.3 Two phase ow: Pure uid 137
wall temperature is 0.86 for all other type of ow patterns it is taken as unity (VDI-
Warmeatlas[157]).
Table A.1 shows the reference factors for the nucleate boiling heat transfer coecient for
R134a and R290.
Table A.1. Values of the reference parameters used in evaluation of the local nucleate boiling
heat transfer coecient.
Refrigerant h
o
q
o
R
ao
d
o
W/m
2
K W/m
2
m m
R134a 3,500 20,000 10
6
0.01
R290 4,000 20,000 10
6
0.01
A.3.2 Kattan et al. [77] correlation
For a stratied-wavy ow pattern or annular ow pattern with a partial dryout the two
phase heat transfer coecient is
h =
dry
h
G
+ (2
dry
)h
wet
2
. (A.33)
The vapor heat transfer coecient h
G
is determined by using the Dittus-Boelter [33]
correlation as
h
G
= 0.023Re
0.8
G
P
0.4
rG
k
G
d
, (A.34)
with Reynold number given as
Re
G
=
m xd
G
, (A.35)
where is the void fraction given by the Rauhani [117] model (equation A.22). The heat
transfer coecient on the wetted portion of the tube is
h
wet
=
3
_
h
3
n
+ h
3
c
. (A.36)
The nucleate boiling heat transfer coecient h
n
is given by the Cooper [27] model as
h
n
= 55p
0.12
r
(0.4343 ln p
r
)
0.55
M
05
q . (A.37)
The convective heat transfer coecient is given by a modied form of the Dittus-Boelter
[33] model as
h
c
= 0.0133Re
0.69
L
P
0.4
rL
k
L
d
. (A.38)
The liquid Reynolds number is given as
Re
L
=
4 m(1 x)
(1 )
G
. (A.39)
where is the liquid lm thickness it is given as
=
d(1 )
2(2
dry
)
, (A.40)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
138 A Heat transfer coecient
where
dry
is
dry
=
strat
( m
wavy
m)
( m
wavy
m
strat
)
, (A.41)
where
strat
is calculated iteratively from equation A.24. The mass ux under a stratied
and wavy ow pattern is
m
strat
=
(226.3)
2
f
L
f
2
G
G
(
L
G
)
L
g cos
0.3164(1 x)
1.75
0.25
L
, (A.42)
and
m
wavy
=
_
16f
3
G
gd
L
G
x
2
2
(1 (2h
L
1)
2
)
0.5
_
2
25h
2
L
(1 x)
F
1
( q)
_
We
Fr
_
F
2
( q)
L
+
1
cos
__
0.5
+ 50 ,
(A.43)
respectively. The parameters f
L
, f
G
, h
L
are dened in Fig.A.2. is the angle of inclination
to the horizontal and
F
1
( q) = 646.0
_
q
q
crit
_
2
+ 64.8
_
q
q
crit
_
; F
2
( q) = 18.8
_
q
q
crit
_
+ 1.023 . (A.44)
The stratied-wavy ow model is also valid for the stratied ow patten with
strat
replacing
dry
and for the annular ow condition with
dry
is set to zero and the lm
thickness is set to (1 )d/4.
A.3.3 Kandlikar [70] correlation
The ow boiling heat transfer coecient for a pure uid is given by Kandlikar [70] as
h = max(h
n
, h
c
) , (A.45)
wher the subscript n and c in equation A.45 refers to the nucleate and convective boiling
respectively. The convective and the nucleate boiling part is given as
h
n
= 0.6683Co
0.2
(1 x)
0.8
h
L0
f(F
rL0
) + 1058.0Bo
0.7
(1 x)
0.8
F
Fl
h
L0
, (A.46)
and
h
c
= 1.136Co
0.9
(1 x)
0.8
h
L0
f(F
rL0
) + 667.2Bo
0.7
(1 x)
0.8
F
Fl
h
L0
, (A.47)
respectively, where Fr
L0
is the liquid Froude number, Bo is the boiling number and Co
is the convection number. These dimensionless groups are dened as
Fr
L0
=
m
L
gd
, Bo =
q
mh
v
, Co =
_
L
_
0.5 _
1 x
x
_
0.8
. (A.48)
The function f(F
rL0
) is dened as
f(F
rL0
) = (25F
rL0
)
0.324
F
rL0
< 0.04 ,
f(F
rL0
) = 1 F
rL0
0.04 ,
where F
Fl
is a uid-surface parameter related to the nucleation characteristic. For all
type of uids owing in a stainless tube it is taken as 1. The single phase heat transfer
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
A.3 Two phase ow: Pure uid 139
coecient h
L0
is obtained from the Petukhov and Popov [114] correlation or Gnielinski
[46] correlation. The Petukhov and Popov [114] correlation is valid in the range of 0.5
P
rL
2000 and 10
4
Re
L0
5 10
6
and it is given as
Nu
L0
=
h
L0
d
k
=
Re
L0
Pr
L
(/2)
1.07 + 12.7(P
2/3
rL
1)(/2)
0.5
. (A.49)
The Gnielinski [46] correlation (equation A.13) is valid in the range of 0.5 P
rL
2000 and 2300 Re
L0
5 10
4
. The friction factor in equation A.49 is given by
equation A.17.
A.3.4 Chen [19] correlation
Chen [19] postulated that the heat transfer coecient is made of two parts: a) a micro-
convective (or nucleate boiling) portion h
n
and b) a macro-convective (or forced convec-
tive) portion h
c
as
h = h
c
F + h
n
S , (A.50)
where h
c
is calculated using the Dittus and Boelter [33] correlation as
h
c
= 0.023
k
L
d
Re
0.8
L
Pr
0.4
L
, (A.51)
where
Re
L
=
(1 x) md
L
, Pr
L
=
c
pL
L
k
L
, (A.52)
The suppression factor for the convection part is
F =
_
_
1 if 1/X
tt
> 0.1
2.35
_
1
X
tt
+ 0.213
_
0.736
if 1/X
tt
0.1
,
and the Martinelli parameter X
tt
is given as
X =
_
1 x
x
_
0.875
_
L
_
0.5
_
G
_
0.125
. (A.53)
The nucleate boiling heat transfer coecient is
h
n
= 0.00122
k
0.79
L
c
0.45
p,L
0.49
L
0.5
0.29
L
0.24
G
h
0.24
V
T
0.24
sat
p
0.75
sat
, (A.54)
where
T
sat
= T
w
T
s
; p
sat
= p(T
w
) p(T
s
); Re
tp
= Re
L
F
1.25
. (A.55)
The suppression factor for the nucleate part is
S =
1
1 + 2.53 10
6
Re
tp
. (A.56)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
140 A Heat transfer coecient
A.3.5 Gungor and Winterton [52] correlation
The Gungor and Winterton [52] correlation is a modied form of the Chen [19] correlation
given by equation A.50 with the nucleate boiling calculated from the Cooper [27] corre-
lation given by equation A.37. The suppression factor for the convection part is dened
as
F =
_
_
(1 + 24, 000Bo
1.16
+ 1.37(1/x
tt
)
0.86
)Fr
(0.12Fr
L
)
L
if Fr < 0.05
1 + 24, 000Bo
1.16
+ 1.37(1/x
tt
)
0.86
if Fr 0.05
,
and the suppression factor for the nucleate part is
S =
_
_
(1 + 0.00000115F
2
Re
L
)
1
Fr
1/2
L
if Fr < 0.05
(1 + 0.00000115F
2
Re
L
)
1
if Fr 0.05
,
The convective boiling part is calculated from the Dittus-Boelter [33] correlation (equation
A.51).
A.3.6 Shah [130] correlation
The Shah [130] correlation is given as
h = max(h
c
, h
n
) , (A.57)
where the subscript n and c in equation A.57 refers to the nucleate and convective boiling
respectively. The convective heat transfer coecient is dened as
h
c
= 1.8h
L
N
0.8
, (A.58)
where
N =
_
_
Co Fr
L
> 0.04
0.38Fr
0.4
L
Co Fr
L
< 0.04
,
where h
L
is calculated using the Dittus-Boelter [33] correlation (equation A.51). The
nucleate boiling heat transfer coecient is calculated as follows
For N > 1
h
n
=
_
_
230h
L
Bo
0.5
Bo > 0.0003
1 + 46h
L
Bo
0.5
Bo < 0.0003
.
For 1 > N > 0.1
h
n
= Fh
L
Bo
0.5
exp(2.74N
0.1
) . (A.59)
For N < 0.1
h
n
= Fh
L
Bo
0.5
exp(2.47N
0.15
) , (A.60)
where
F =
_
_
14.7 Bo > 0.0011
15.43 Bo < 0.0011
.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
A.3 Two phase ow: Pure uid 141
A.3.7 Schrock and Grossman [129] correlation
A very simple correlation is given by Schrock and Grossman [129] as
h = 1.91h
L
_
10
4
Bo + 1.5
_
1
X
tt
_
2/3
_
0.6
, (A.61)
where h
L
is calculated using Dittus-Boelter [33] correlation equation A.51.
A.3.8 Dembi et al. [30] correlation
The Dembi et al. [30] correlation is based on the asymptotic model given by equation
A.11 with the nucleate and convection part given as
h
n
= 23388.5
k
L
d
_
q
G
h
V
_
0.64
_
gd
h
V
_
0.27
_
m
2
d
L
h
V
_
0.14
, (A.62)
and
h
c
= 0.115
k
L
d
_
x
4
(1 x)
2
_
0.11
_
m
2
h
V
L
g
_
0.14
P
0.27
rL
, (A.63)
respectively. The parameter is dened as
= 0.36 10
3
p
1.4
r
. (A.64)
A.3.9 Klimenko [84] correlation
The Klimenko [84] correlation is based on the asymptotic model given by equation A.11
with the convection part given by the Dittus-Boelter [33] correlation equation A.51 and
the nucleate boiling is
h
n
=
_
h
n1
N
CB
< 1.6 10
4
h
n2
N
CB
> 1.6 10
4
,
where
h
n1
= 7.4 10
3
_
k
w
k
L
_
0.15
Pe
0.6
K
0.5
p
Pr
1/3
L
, (A.65)
h
n2
= 0.087
k
L
b
_
k
w
k
L
_
0.09
Re
0.6
m
_
L
_
0.2
Pr
1/6
L
, (A.66)
Pe =
_
qb
h
V
G
a
L
_
, Kp =
p
_
g(
L
G
)
, b =
2
g(
L
G
)
, (A.67)
Re
m
=
w
m
b
L
, w
m
=
m
L
_
1 + x
_
G
1
__
, Re
=
qb
h
V
L
, N
CB
=
Re
m
Re
G
_
.
(A.68)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
142 A Heat transfer coecient
A.3.10 Jung et al. [64] correlation
The Jung et al. [64] correlation is a modied form of the Chen [19] correlation. The
convection heat transfer coecient is calculated using the Dittus-Boelter [33] correlation
(equation A.51) and the nucleate part is calculated from the Stephan and Abdelsalm in
VDI-Warmeatlas [157] correlation as
h
n
= 207
k
L
b.d
_
q(b.d)
k
L
T
s
_
0.745
_
L
_
0.581
P
0.533
rL
, (A.69)
where
(b.d) = 0.511
_
2
g(
L
G
)
_
0.5
, (A.70)
F = 2.37
_
0.29 +
1
X
tt
_
, (A.71)
S =
_
_
4048X
1.22
tt
Bo
1.13
Xtt < 1
2.0 0.1X
0.28
tt
Bo
0.33
1 X
tt
5
.
A.4 Two phase ow: Mixture
A.4.1 Steiner [140] correlation
Steiner [140] has extended his pure component asymptotic model to mixture. The nucleate
part of the heat transfer coecient is suppressed using the Schl under [126] suppression
factor for the nucleate boiling. The Schl under [126] suppression factor is based on the
heat and mass transfer laws it is dened as
F
n
=
_
1 +
h
id,n
q
(T
b,k
T
b,j
)( y
j
x
j
)
_
1 exp
B
o
q
L
h
V
L
__
, (A.72)
where T
b
is the saturated (boiling) temperature of the pure component, the index j and
k stands for the more volatile and less volatile component respectively.
L
/B
0
= 5 10
5
is the mass transfer coecient. The ideal nucleate boiling heat transfer coecient for a
mixture h
id,n
is calculated from the heat transfer coecient of pure components as
h
id,n
=
_
x
i
h
i,n
_
1
, (A.73)
and B
o
/
L
= 5.10
3
and
L
and h
V
is the ideal density and enthalpy of evaporation of
the mixture respectively. x and y is the liquid and vapor mole fraction of the more volatile
component respectively.
The same approach applies also to the convective part for the liquid-liquid immiscible
mixture. That is to say for a liquid-liquid miscible mixture the convective suppression
factor made analogous to that for the nucleate boiling heat transfer coecient as
F
c
=
_
1 +
h
id,c
q
(T
b,k
T
b,j
)( y
j
x
j
)
_
1 exp
B
o
q
L
h
V
L
__
. (A.74)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
A.4 Two phase ow: Mixture 143
A.4.2 Kandlikar [71] correlation
Kandlikar [71] has extended his pure component correlation (Kandlikar [70]) to mixtures
as
Region I: Near-azeotropic region
h = max(h
n
, h
c
) , (A.75)
where h
n
and h
c
is obtained from equation A.77 and equation A.47 respectively
using the mixture properties.
Region II: Moderate diusion-induced suppression region
h = h
c
, (A.76)
where h
c
is given by equation A.77 with the properties of the mixture.
Region III: Severe diusion-induced suppression region: 0.03< V
1
< 0.2 and Bo
1E
4
; V
1
0.2
h = 1.136Co
0.9
(1 x)
0.8
h
L0
f(F
rL0
) + 667.2Bo
0.7
(1 x)
0.8
F
Fl
h
L0
F
D
, (A.77)
where
V
1
=
_
_
c
pL
h
V
__
a
D
12
_
0.5
| y x|
_
dT
d x
__
, (A.78)
F
D
=
0.678
1 +V
1
. (A.79)
A.4.3 Bennett and Chen [8] correlation
Bennett and Chen [8] has extended the Chen [19] correlation (equation A.50) for mix-
ture. Here both the convective and the nucleate parts are suppressed. The convection
part which is calculated for the original Chen [19] correlation with mixture properties is
suppressed using the following suppression factor
F
c
=
T
w
T
ph
T
w
T
s
, (A.80)
where T
w
, T
ph
, and T
s
is the wall, equilibrium temperature and saturation temperature
respectively. The nucleate part is also calculated using the original Chen [19] model for
the pure substance with mixture properties. It suppressed using the the suppression factor
given by equation A.79.
A.4.4 Palen [111] correlation
Palen [111] has extended the original Chen [19] correlation for pure component (equation
A.50) to mixture similar to the Bennett and Chen [8] correlation. However, only the
nucleate part is suppressed using the following suppression factor
F
d
= exp(0.027T
bp
) , (A.81)
where T
bp
is dierence between the dew and bubble point temperature of the mixture.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
144 A Heat transfer coecient
A.4.5 Jung et al. [64] correlation
Jung et al. [64] have extended their pure substance correlation to the mixture. The nu-
cleate boiling heat transfer coecient is replaced by the ideal one given by equation A.73.
The convective part is suppressed using the following suppression factor
F
c
= 1.0 0 35| y
1
x
1
|
1.56
. (A.82)
For the nucleate part the following suppression factor is employed
F
n
=
1
{[1 + (b
2
+ b
3
)(1 + b
4
)](1 + b
5
)}
2
, (A.83)
where
b
2
= (1 x
1
) ln
_
1.01 x
1
1.01 y
1
_
+ x
1
ln
_
x
1
y
1
_
+| y
1
x
1
|
1.5
, (A.84)
b
3
=
_
_
0 x
1
0.01
_
x
1
y
1
_
0.1
1 x
1
< 0.01
,
b
4
= 152
_
p
p
c,1
_
0.66
, (A.85)
b
5
= 0.92| y
1
x
1
|
0.001
_
p
p
c,1
_
0.66
, (A.86)
and
x
1
y
1
= 1 for x
1
= y
1
= 0 ,
x
1
and y
1
is the liquid and vapor mole fraction of the more volatile component respec-
tively. p and p
c,1
is system pressure and critical pressure of the more volatile component
respectively.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
145
B Pressure drop
B.1 Single phase
The pressure drop due to friction exists because of the shear stress between the uid and
the tube wall. Estimation of the friction pressure drop is somewhat more complex and
various approaches have been taken, for example the frictional pressure gradient is given
as
_
dp
dz
_
f
=
4
o
d
=
4f m
2
2d
, (B.1)
where m is the mass ux in kg/m
2
s and f is the friction factor calculated using a Blasius-
type model as
f =
_
_
0.3164
Re
0.25
Re 2320
64
Re
Re < 2320 .
Integration of equation B.1 yields
p =
4f m
2
2
L
d
, (B.2)
B.2 Two phase
In ow boiling, the temperature drops in the direction of ow as a result of the pressure
drop. This results in a change in the driving force (temperature dierence) for the heat
transfer along the ow path. Thus beside the heat transfer coecient, knowledge of the
pressure drop is of paramount importance in the design of the evaporator. In the present
work the pressure drop is measured simultaneously with the heat transfer coecient along
the test section.
The momentum balance implies that the two phase pressure gradient is composed of three
components as
dp
dz
=
_
dp
dz
_
f
+
_
dp
dz
_
a
+
_
dp
dz
_
h
, (B.3)
where dp/dz, (dp/dz)
f
, (dp/dz)
a
and (dp/dz)
h
is the total, friction, acceleration and
hydrostatic pressure gradient respectively. For a horizontal tube the hydrostatic pressure
gradient diminishes. The acceleration pressure drop is caused by the change in momentum
in both the liquid and vapor phases. The change in the momentum stems from the change
in the velocity of the two phases, which is brought about by the added (or withdrawn)
heat to/from the test section. For the case of adiabatic ow the acceleration pressure drop
diminishes for p
a
/p
s
0 (Baehr and Stephan [3]), where p
s
is the saturation pressure.
There exist in the literature a number of approaches for modelling the change in the static
pressure drop due to acceleration. The most widely accepted models include homogenous
or separated ow models. The separated ow model is also widely known as the het-
erogenous model. In the homogenous model the static pressure drop due to acceleration
is
_
dp
dz
_
a
= m
2
d
dz
_
x
_
1
G
_
+
1
L
_
. (B.4)
The energy balance in a small unit length dz along the test tube yields
d x
dz
=
4 q
mh
v
d
. (B.5)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
146 B Pressure drop
Substitution of equation B.5 into equation B.4 yields the pressure drop due to acceleration
as
p
a
=
4 q m
dh
v
G
_
1
G
L
_
L . (B.6)
In the separated ow model the static pressure drop due to acceleration can be derived
from the momentum balance as
_
dp
dz
_
a
= m
2
d
dz
_
x
2
G
+
(1 x)
2
(1 )
L
_
. (B.7)
Integration of equation B.7 between the inlet i and outlet o of the test section yields
p
a
= (p
o
p
i
)
a
= m
2
_
x
2
2
G,o
+
(1 x
o
)
2
(1
o
)
L,o
x
2
i
G,i
(1 x
i
)
2
(1
i
)
L,i
_
. (B.8)
The void fraction may be obtained using the Rauhani [117] model which is given as:
=
x
G
_
(1 + 0.12(1 x))
_
x
G
+
1 x
L
_
+
1.18(1 x)[g(
L
G
)]
1/4
m
1/2
L
_
1
, (B.9)
where
L
and
G
is the liquid and vapor density respectively, which are calculated from the
fundamental equation of state of Tillner-Roth and Baehr [152] for R134a. g is acceleration
due to gravity, is the surface tension, m is the mass ux and x is the quality. The surface
tension is calculated using the method of Lucus [92] given in VDI-Warmeatlas [157].
The pressure drop due to friction exists because of the shear stress between the uid and
the tube wall. Estimation of the friction pressure drop is somewhat more complex and
various approaches have been taken, for example in homogenous or separated ow models.
In the homogenous model the frictional pressure gradient is given as
_
dp
dz
_
f
=
4
o
d
=
2 m
2
d
H
, (B.10)
where is the two phase friction factor calculated by a Blasius-type model as
=
_
_
0.3164
Re
0.25
Re 2320
64
Re
Re < 2320 .
and the homogenous density
H
is given as
1
H
=
1 x
L
+
x
G
. (B.11)
The two phase Reynolds number Re is
Re =
md
TP
, (B.12)
where
TP
is a two-phase viscosity. A variety of methods have been proposed to calculate
the two phase viscosity, a commonly used one being that proposed by McAdams et al. [95]
1
TP
=
1 x
L
+
x
G
, (B.13)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
B.2 Two phase 147
where
L
and
G
are the liquid and vapor viscosity.
In the separated ow model the two phase frictional pressure drop is related to that for
single phase as
_
dp
dz
_
f
=
_
dP
dz
_
f,L/G
G/L
, (B.14)
where is the two phase multiplier. There exist a number of correlations for the prediction
of . These include Friedel [42], Chishlom [22] and Lockhart and the Martinelli [91] model.
These models are presented in Appendix B. There exists a number of correlations for the
prediction of the two phase multiplier of the separated ow model. These models are
presented in the following subsections.
B.2.1 Friedel [42] model
L0
= E +
3.24FH
Fr
0.045
We
0.035
, (B.15)
where
E = (1 x)
2
+ x
2
L
f
G0
G
f
L0
, (B.16)
F = x
0.78
(1 x)
0.24
, (B.17)
H =
_
G
_
0.91
_
L
_
0.19
_
1
G
L
_
0.7
, (B.18)
Fr =
m
2
gd
2
H
, (B.19)
We =
m
2
d
H
, (B.20)
d is tube diameter, is the surface tension and
H
is the homogenous density given by
equation B.11. f
G0
and f
L0
are the friction factors dened by a Blasius-type model as
f
L0/G0
=
0.079
Re
0.25
L0/G0
, (B.21)
where Re = md/. The range of the validity of the Friedel [42] model is
L
/
G
< 1000
B.2.2 Lockhart and Martinelli [91] model
In the Lockhart and Martinelli [91] model the two phase friction multiplier is
2
L
= 1 +
C
X
+
1
X
2
, (B.22)
2
G
= 1 + C.X +X
2
, (B.23)
where X is the Martinelli parameter and the value of the coecient C is given in Table B.1.
The range of the applicability of the Lockhart and Martinelli [91] correlation is
L
/
G
>1000
and m <100 kg/m
2
s.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
148 B Pressure drop
Table B.1. Value of C for the Lockhart and Martinelli [91] correlation.
Liquid Gas Subscript C
Turbulent Turbulent tt 20
Viscous Turbulent vt 12
Turbulent Viscous tv 10
Viscous Viscous vv 05
B.2.3 Chisholm [22] model
In the Chisholm [22] model the two phase friction multiplier is
L0
= 1 + (Y
2
1)
_
B x
(2/n1)
(1 x)
(2/n1)
+ x
1n
_
, (B.24)
where
Y
2
=
(dp
f
/dz)
G0
(dp
f
/dz)
L0
, (B.25)
n is 0.25 for a Blasius model. The parameter B is given by
B =
55
m
1/2
0 < Y < 9.5 , (B.26)
B =
520
Y m
1/2
9.5 < Y < 28 , (B.27)
B =
15000
Y
2
m
1/2
28 < Y . (B.28)
The range of the validity of the Chisholm [22] correlation is
L
/
G
> 1000 and m > 100
kg/m
2
s.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
149
C Physical properties
The uid physical properties required for heat exchanger design are divided in thermo-
dynamic and trasport properties. The transport properties include viscosity, thermal
conductivity, surface tension and diusion coecient are generally calculated from the
existing correlations (Pery and Coulson). The thermodynamic properties include dem-
sity, specic heat temperature, pressure (vapor), enthalpy, latent heat of evaporation.
Beside the uid properties the thermal conductivity of the material is necessary for the
evaluation of heat transfer coecient. The thermodynamic properties are evaluated using
critical tables.
C.1 Physical properties: Pure uid
C.1.1 Specic heat
The specic heat of the ideal gas is given in as
Cp = CPV APA + (CPV APB)T + (CPV APC)T
2
+ (CPV APD)T
3
(C.1)
Where T is in K and CPVAPA, CPVAPB, CPVAPC, CPVAPD are constant in ideal
gas heat capacity. These constant are given in Appendix A for organic and inorganic
compounds.
C.1.2 Vapor pressure
The vapor pressure is generally predicted using Antonie equation as
ln p = ANTA
ANTB
T + ANTC
(C.2)
where T is in K and ANTA, ANTB,ANTC are Anonie equation constant. These constant
are given in Appendix D for organic and inorganic compounds.
C.1.3 Liquid viscosity
The liquid viscosity is given as:
log = V ISA
_
1
T
1
V ISB
_
(C.3)
where VISA, VISB are constants in the liquid viscosity equation. These constant are
given in Appendix D for organic and inorganic compounds.
C.1.4 Vapor dynamic viscosity VDI-Warmeatlas [157]
Lucas and Luckas [92] in VDI-Warmeatlas [157] have recommended the following proce-
dure for the calculation of the vapor viscosity.
= ()
r
F
p
F
Q
1
, (C.4)
for T
r
1 and p
r
p
s
/p
c
()
r
= 0.600 + 0.760p
r
+ (6.990p
r
0.6)(1 T
r
) , (C.5)
with
= 3.262 + 14.98p
5.508
r
and = 1.390 + 5.746p
r
, (C.6)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
150 C Physical properties
for 1 T
r
40 and 0 p
r
100
()
r
= (
o
)
_
1 +
Ap
E
r
Bp
F
r
+ (1 + Cp
D
r
)
1
_
, (C.7)
where
o
is the low pressure viscosity given as
o
= [0.807T
0.618
r
0.357 exp(0.449T
r
) + 0.340 exp(4.058T
r
) + 0.018]F
o
p
F
o
Q
, (C.8)
and is given as
=
[T
c
]
1/6
[R]
1/6
[N
a
]
1/3
[M]
1/2
[p
c
]
2/3
, (C.9)
where N
a
is the Avagadro number in kmol. The coecients of equation C.7 are given as
A =
a
1
T
r
exp(a
2
T
r
) , (C.10)
B = A(b
1
T
r
b
2
) , (C.11)
C =
c
1
T
r
exp(c
2
T
r
) , (C.12)
D =
d
1
T
r
exp(d
2
T
r
) , (C.13)
E = 1.3088 , (C.14)
F = f
1
exp(f
2
T
r
) . (C.15)
The coecients a, b, c, d, e, and f are given in Table C.1
Table C.1. Coecients of the correlation used for the prediction of the vapor dynamic viscosity.
a
1
1.245.10
3
a
2
5.1726 c
1
0.4489 c
2
3.0578 -0.3286
b
1
1.6553 b
2
1.2723 d
1
1.7368 d
2
2.2310 -37.7332
f
1
0.9425 f
2
0.1853 0.4489 -7.6351
F
p
= 1 + (F
o
p
1)
_
()
r
_
3
, (C.16)
and
F
Q
= 1 + (F
o
Q
1)
_
()
r
_
1
0.007
_
ln
_
()
r
__
4
, (C.17)
where F
o
p
and F
o
Q
is low-pressure polarity and quantum factors respectively. These factors
are
F
o
p
= 1 , 0
r
< 0.022 , (C.18)
F
o
p
= 1 + 30.55(0.292 Z
c
)
1.7
, 0.022
r
< 0.075 , (C.19)
F
o
p
= 1 + 30.55(0.292 Z
c
)
1.7
(|0.96 + 0.1(T
r
0.7)|) , 0.075
r
, (C.20)
where Z
c
is the critical compressibility factor and F
o
Q
= 1.0 for all substances other than
He, H
2
and D
2
. The reduced dipole moment
r
is given as
r
=
2
p
c
(kT
c
)
2
, (C.21)
where the dipole moment for the gases is given in VDI-Warmeatlas [157]
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
C.1 Physical properties: Pure uid 151
C.1.5 Dynamic viscosity of Fenghour et al. [40]
The functional form of the liquid and vapor viscosity of ammonia as given by Fenghour
et al. [40] is
=
o
(T) +
1
(T) +
2
(, T) , (C.22)
The rst term of the expansion is the dilute gas term which is given as
o
(T) = 100
_
0.021357
0.2957
2
_
(
MT)
1/2
exp()
, (C.23)
where
M is the molecular weight in g/mol, T is the temperature in K. The collision
integral is dened as
(T) =
_
C(1) + C(2) log
_
kT
_
+
4
n=3
C(n)
_
log
_
kT
__
n
_
, (C.24)
where /k=386 K and the value of the coecient C is given in table C.2.
Table C.2. Coecients for the Collision integral (equation C.24).
C(1) 4.9931822 C(2) -0.61122364 C(3) 0.18535124 C(4) -0.1116094
The second term of equation C.22 represents the contribution of the moderately dense
uid
1
(T) = F
v
(T)
o
(T) , (C.25)
where
F
v
(T) = C
_
_
A(1) +
13
i=2
A(i)
_
log
_
kT
__
(i1)
2
_
_
, (C.26)
where C=0.6022137/0.2957
3
and the value of the coecient A is given in table C.3
Table C.3. Coecients of equation C.26.
i A i A
1 -0.1799949610
1
2 0.46669262110
2
3 -0.5346079410
3
4 0.3360407410
4
5 -0.1301916410
5
6 0.3341423010
5
7 -0.5871174310
5
8 0.7142668610
5
9 -0.5983401210
5
10 0.3365274110
5
11 -0.1202735010
5
12 0.2434820510
4
13 -0.12080795710
3
The third term in the viscosity equation C.22 is the contribution of the dense gas
2
(, T) =
3
i=1
F(i, T)
i+1
, (C.27)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
152 C Physical properties
where
F(i, T) =
_
_
1 0.219664285
_
kT
_
2
0.83651107 10
1
_
kT
_
4
2 0.17366936 10
2
0.83651107 10
2
_
kT
_
3 0.167668649 10
3
_
kT
_
2
0.149710093 10
3
_
kT
_
3
+
0.77012274 10
4
_
kT
_
4
The Fenghour et al. [40] correlation for the vapor viscosity of ammonia has an uncertainty
of 2% in the temperature range of T < T
c
.
C.1.6 Surface tension
Lucas and Luckas [92] in VDI-Warmeatlas [157] have recommended the following corre-
lation for the calculation of the surface tension
= p
2/3
c
T
1/3
c
_
1 T
r
a
_
m
b , (C.28)
where the reduced pressure and temperature are dened as
p
r
=
p
p
c
, T
r
=
T
T
c
, , (C.29)
respectively.
For a polar uid like R134a the following quantities are valid
a = 1 , (C.30)
b = 0.1574 + 0.359 1.769X 13.69X
2
0.510
2
+ 1.298X , (C.31)
m = 1.210 + 0.5385 14.61X 32.07X
2
1.656
2
+ 22, 03X , (C.32)
X = lgp
sr
(T
r
= 0.6) + 1.70 + 1.552 . (C.33)
where is the acentric factor and it is given by Pitzer in VDI-Warmeatlas [157] as The
surface tension given by equation C.28 is in 10
5
N/cm. Its level of uncertainty as given
by Reid et al. [118] is 1.2 % in the range of the reduced temperature of 0.56 T
r
0.63.
C.1.7 Thermal conductivity for liquids
k = 3.65 10
5
C
p
_
M
_
1/3
. (C.34)
where k thermal conductivity W/moC, M is the molecular mass, C
p
speic heat capacity
(kJ/kg oC), density (kg/m
3
)
C.1.8 Thermal conductivity for gases
k =
_
C
p
+
10.4
M
_
. (C.35)
where k thermal conductivity W/m
o
C, M is the molecular mass, C
p
specic heat capacity
(kJ/kg
o
C), viscosity in (mNs/m
2
)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
C.2 Physical properties: Mixture 153
C.1.9 Specic enthalpy
For the vapor phase, the deviation of the specic enthalpy from the ideal state can be
illustrated using Redlich-Kwong equation written as
z
3
+z
2
+ z(B
2
+ B A) = 0 . (C.36)
where z is the compressibilty factor dened as
z =
pv
RT
. (C.37)
and
A =
aP
R
2
T
2.5
, B =
bp
RT
. (C.38)
h = h
o
+ RT +
_
v
0
__
T
dP
R
2
T
2.5
dT
_
p
_
dv . (C.39)
C.2 Physical properties: Mixture
C.2.1 Liquid dynamic viscosity of mixtures
For a liquid mixture which contains one or more polar constituents Reid et al. [118]
recommended the following model for the calculation of the mixture liquid viscosity
ln
m
=
n
i=1
x
i
. ln
L,i
+ 2. x
1
. x
2
.G
12
, (C.40)
where x
i
is the mole fraction of the component i,
L,i
is the viscosity of the component i
in kg/ms and G
12
is an adjustable parameter normally obtained from experimental data.
For a polar-nonpolar mixture G
12
= -0.22. The Reid et al. [118] model give the thermal
conductivity with a mean error of less then 5%.
C.2.2 Vapor dynamic viscosity of mixtures
The viscosity of a gas mixture can be approximated by using the principle of the kinetic
theory (Reid et al. [118]) as
m
=
o
m
+ , (C.41)
where
o
m
is the mixture gas viscosity at a low pressure and is a correction factor for
the high pressure viscosity
o
m
=
n
i=1
y
i
G,i
n
j=1
y
i
ij
, (C.42)
where y
i
is the mole fraction of the component i and
i
is the viscosity of the pure
component i.
ij
is a parameter which may be estimated as
ij
=
_
1 + (
G,i
/
G,j
)
0.5
(
M
j
/
M
i
)
0.25
_
2
[8(1 +
M
i
/
M
j
)]
0.5
, (C.43)
ji
=
G,j
G,i
M
j
M
i
ij
. (C.44)
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
154 C Physical properties
The high pressure correction term is estimated as
=
0.497.10
6
_
exp(1.439
r,m
) exp(1.111
1.858
r,m
)
_
T
1/6
c,m
M
0.5
m
p
2/3
c,m
. (C.45)
The pseudo critical properties of the mixture are calculated as
T
c,m
=
j=1
y
j
T
c,j
,
c,m
=
j
y
j
c,j
,
Z
c,j
=
p
c,j
c,j
RT
c,j
,
Z
m
=
j
y
j
Z
c,j
, (C.46)
M
m
=
j=1
y
j
M
j
,
c,m
=
M
m
/1000
c,m
,
r,m
=
m
c,m
, p
c,m
=
RT
c,m
Z
c,m
c,m
, (C.47)
where T is in K, p is in Mpa,
c,m
is in m
3
/kmol,
r,m
is in kg/m
3
, M is in g/mol and
m
is in kg/ms. The error associated with this model is seldom exceeded 3 to 4% (Perry and
Green [112]).
C.2.3 Liquid thermal conductivity of mixtures
Reid et al. [118] have recommended a Filippov-like model for the prediction of the thermal
conductivity of a liquid mixture as
m
=
2
i=1
X
i
L,i
0.72X
1
X
2
|
L,2
L,1
| , (C.48)
where
X
1
and
X
2
is the weight fraction of the component 1 and 2 respectively and
1
and
2
is the thermal conductivity of the component 1 and 2 in W/mK respectively.
C.2.4 Vapor thermal conductivity of mixtures
The thermal conductivity of a low-pressure gas mixture can be determined from the
relationship given by Reid et al. [118]
G,m
=
n
i=1
y
i
G,i
n
j=1
y
i
A
ij
, (C.49)
where
G,m
is the low-pressure gas mixture thermal conductivity,
G,i
is the low-pressure
thermal conductivity of the pure component i. For a binary mixture of two non-polar
gases or a non-polar and a polar gas, A
ij
may be calculated by the model given by Perry
and Green [112] as
A
ij
=
_
1 + (
tr,i
/
tr,j
)
0.5
(
M
j
/
M
i
)
0.25
_
2
[8(1 +
M
i
/
M
j
)]
0.5
, (C.50)
with
tr,i
tr,j
=
j
i
exp(0.0464T
r,i
) exp(0.2412T
r,i
)
exp(0.0464T
r,j
) exp(0.2412T
r,j
)
, (C.51)
where
M is the molecular weight and is dened as
i
= 210
_
T
c,i
M
3
i
P
4
ci
_
(1/6)
, (C.52)
where T is in K, p is in bar,
M is in g/mol and is in W/mK. This model yields an error
of less than 5% in the prediction of the thermal conductivity of the gas mixture.
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com
C.3 Software packages 155
C.2.5 Surface tension of mixtures
Lucas and Luckas [92] in VDI-Warmeatlas [157] recommended the following method for
calculation of the mixture surface tension
m
= p
2/3
c,m
T
1/3
c,m
_
1 t
r,m
a
m
_
n
m
b
m
, (C.53)
where
b
i
= 0.1196.
_
1 +
T
s,ri
ln(p
c,m
/1.01325)
1 T
s,ri
_
, b
m
=
x
i
b
i
, (C.54)
a
m
= 1, n
m
= 11/9, T
c,m
=
j=1
x
i
T
c,j
,
c,m
=
j
x
j
c,j
,
Z
c,j
=
p
c,j
c,j
RT
c,j
, (C.55)
Z
m
=
j
x
j
Z
c,j
, p
c,m
=
RT
c,m
Z
c,m
c,m
, T
s,ri
=
T
b,i
T
c,i
, (C.56)
where T
b,i
=T (p=1.01325 bar) is the normal boiling point temperature of the pure com-
ponent i. T is in K, p is in bar and is in N/m. The Lucas and Luckas correlation yields
an error of <5%.
C.3 Software packages
There exists a number of software packages for the prediction of thermodynamic and
transport properties. These include:
1. ASPEN Plus (
2. CHEMCAD
3. SUPERPRO
4. REFPROP
Dr. Ali A. Rabah, Dept of Chemeng, U of K, Email : rabahss@hotamil.com