This action might not be possible to undo. Are you sure you want to continue?
ShellandTube Heat Exchangers
R. Shankar Subramanian
Shellandtube heat exchangers are used widely in the chemical process industries, especially in
refineries, because of the numerous advantages they offer over other types of heat exchangers. A
lot of information is available regarding their design and construction. The present notes are
intended only to serve as a brief introduction.
For detailed information about analyzing and designing shellandtube heat exchangers, consult
“The Chemical Engineers’ Handbook” (http://www.knovel.com/knovel2/Toc.jsp?BookID=48 )
(Chapter 11) or any of a variety of sources on heat exchanger design. Mechanical standards for
shellandtube heat exchangers are set by TEMA (Tubular Exchangers Manufacturers
Association) and these supplement the ASME code for such heat exchangers. API (American
Petroleum Institute) Standard 660 supplements both of these standards, and chemical and
petroleum companies also have their own internal standards in addition.
Advantages
Here are the main advantages of shellandtube heat exchangers (Thanks to Professor Ross
Taylor for this list).
1. Condensation or boiling heat transfer can be accommodated in either the tubes or the shell,
and the orientation can be horizontal or vertical. You may want to check out the orientation of
the heat exchanger in our laboratory. Of course, single phases can be handled as well.
2. The pressures and pressure drops can be varied over a wide range.
3. Thermal stresses can be accommodated inexpensively.
4. There is substantial flexibility regarding materials of construction to accommodate corrosion
and other concerns. The shell and the tubes can be made of different materials.
5. Extended heat transfer surfaces (fins) can be used to enhance heat transfer.
6. Cleaning and repair are relatively straightforward, because the equipment can be dismantled
for this purpose.
Basic considerations
The tube side is used for the fluid that is more likely to foul the walls, or more corrosive, or for
the fluid with the higher pressure (less costly). Cleaning of the inside of the tubes is easier than
cleaning the outside. When a gas or vapor is used as a heat exchange fluid, it is typically
introduced on the shell side. Also, high viscosity liquids, for which the pressure drop for flow
through the tubes might be prohibitively large, can be introduced on the shell side.
2
The most common material of construction is carbon steel. Other materials such as stainless
steel or copper are used when needed, and the choice is dictated by corrosion concerns as well as
mechanical strength requirements. Expansion joints are used to accommodate differential
thermal expansion of dissimilar materials.
Heat transfer aspects
The starting point of any heat transfer calculation is the overall energy balance and the rate
equation. Assuming only sensible heat is transferred, we can write the heat duty Q as follows.
( ) ( )
, , , , , , hot p hot hot in hot out cold p cold cold out cold in
Q m C T T m C T T = − = −
lm
Q UA F T = ∆
The various symbols in these equations have their usual meanings. The new symbol F stands
for a correction factor that must be used with the log mean temperature difference for a
countercurrent heat exchanger to accommodate the fact that the flow of the two streams here
is more complicated than simple countercurrent or cocurrent flow. Consider the simplest
possible shellandtube heat exchanger, called 11, which means that there is a single shell “pass”
and a single tube “pass.” The sketch schematically illustrates this concept in plan view. Note
that the contact is not really countercurrent, because the shell fluid flows across the bank of
tubes, and there are baffles on the shell side to assure that the fluid does not bypass the tube
bank. The entire bundle of tubes (typically in the hundreds) is illustrated by a single line in the
sketch. The baffle cuts are aligned vertically to permit dirt particles settling out of the shell side
fluid to be washed away.
1
T
2
T
1
t
2
t
Baffle
3
The convention in shellandtube heat exchangers is as follows:
1
: T inlet temperature of the shellside (or hot) fluid
2
: T exit temperature of the shellside (or hot) fluid
1
: t inlet temperature of the tubeside (or cold) fluid
2
: t exit temperature of the tubeside (or cold) fluid
Thus,
( ) ( )
1 2 2 1
1 2
2 1
ln
lm
T t T t
T
T t
T t
− − −
∆ =
( −
(
−
¸ ¸
The fraction of the circular area that is open in a baffle is identified by a “percentage cut” and we
refer to the types of baffles shown as “segmented” baffles. For the shell side, in evaluating the
Reynolds number, we must find the crossflow velocity across a bundle of tubes that occurs
between a pair of baffles, and determine the value of this velocity where the space for the flow of
the fluid is the smallest (maximum velocity). For the length scale, the tube outside diameter is
employed.
Most shellandtube heat exchangers have multiple “passes” to enhance the heat transfer. Here
is an example of a 12 (1 shell pass and 2 tube passes) heat exchanger.
1
T
2
T
1
t
2
t
Baffle
As you can see, in a 12 heat exchanger, the tubeside fluid flows the entire length of the shell,
turns around and flows all the way back. It is possible to have more than two tube passes.
Multiple shell passes also are possible, but involve fabrication that is more complex and is
usually avoided, if possible.
Correction factors to be used in the rate equation have been worked out by analysis, subject to a
set of simplifying assumptions, for a variety of situations. In the olden days, the formulae for
4
them were considered too cumbersome to use. Therefore graphs were prepared plotting
( ) , F P R , where
2 1
1 1
t t
P
T t
−
=
−
and
1 2
2 1
T T
R
t t
−
=
−
are parameters on which F depends. Figures C4.a
d in Appendix C of the textbook by Mills display such graphs. Nowadays, one can compute these
factors quickly with a pocket calculator. Given next are the two common factors.
2
1 2
2
2
1 1
ln
1 1
1
ln
1
R P
R PR
F
A R
A R
−
(
+ −  
(

− −
\ .
(
¸ ¸
=
(
+ +
(
− + (
¸ ¸
( )
2
2 4
2
2
1 1
ln
2 1 1
1
ln
1
R P
R PR
F
A B R
A B R
−
(
+ −  
(

− −
\ .
(
¸ ¸
=
(
+ + +
(
+ − + (
¸ ¸
where ( )( )
2 2
1 , 1 1 A R B P PR
P P
= − − = − −
The first and second subscripts on the factor F correspond to the number of shell and tube
passes, respectively. The simplifying assumptions mentioned in the previous paragraph, given
in Perry’s Handbook, are as follows.
1. The heat exchanger is at steady state.
2. The specific heat of each stream remains constant throughout the exchanger.
3. The overall heat transfer coefficient U is constant.
4. All elements of a given fluid stream experience the same thermal history as they pass through
the heat exchanger (see footnote in Perry for a discussion regarding the violation of this
assumption in shellandtube heat exchangers).
5. Heat losses are negligible.
The formula given above for
1 2
F
−
also applies for one shell pass and 2, 4, (or any multiple of 2)
tube passes. Likewise, the formula for
2 4
F
−
also applies for two shell passes and 4, 8, (or any
multiple of 4) tube passes.
In designing heat exchangers, one should avoid the steep portion of the curves of F versus P ,
because small errors in estimating P can cause large changes in the value of F . A misleading
rule of thumb is that 0.8 F ≥ , but the correct idea is that the region of steep falloff in the curves
should be avoided.
5
Heat Transfer Coefficients
The evaluation of the overall heat transfer coefficient is an important part of the thermal design
and analysis of a heat exchanger. You’ll find several tables of typical overall heat transfer
coefficients in shellandtube heat exchangers in Chapter 11 of Perry’s Handbook. The following
generic result can be written for the overall heat transfer coefficient
o
U based on the outside
surface area of the tubes, which is the heat transfer surface.
,0 ,
1 1 1
o o
f f i
o o lm i i
A A r
R R
U h k A h A
    ∆
= + + + +
 
\ . \ .
In the above equation,
o
h is the heat transfer coefficient for the fluid flowing in the shell,
i
h is
the heat transfer coefficient for the fluid flowing through the tubes,
i
A and
o
A are the inside and
outside surface areas of a tube, respectively, and
lm
A is their log mean. The fouling resistances
on a unit area basis are
,0 f
R for the shell side, and
, f i
R for the tube side. Accumulated
information on fouling resistances can be found in the Standards published by TEMA.
The inside heat transfer coefficient
i
h can be evaluated using the standard approach for
predicting heat transfer in flow through tubes, including applying a viscosity correction where
possible. Typically, turbulent flow can be expected, and a good design would aim to arrange for
turbulent flow, because of the substantial enhancement in heat transfer provided by eddy
transport. Predicting the shellside heat transfer coefficient
o
h is more involved, because the
flow passage is not simple, even in the absence of baffles. The presence of baffles needs to be
taken into account in calculating the fluid velocity across the tube bank. Heat transfer
correlations for flow through tube banks are used, such as those given in the book by Holman
(1). These correlations assume flow normal to the long axes of a set of tubes placed in a
geometrical array. The correlation given in Holman’s book is
1/ 3
Re Pr
n o o
h D
Nu C
k
= =
The Reynolds number
max
Re
o
D V ρ
µ
= , where
o
D is the outside diameter of a tube.
max
V is the
“maximum” velocity of the fluid through the tube bank. To find it, first, the crossflow area
must be evaluated. This is given as
clearance
Cross flow area Shell ID Baffle spacing
pitch
= × ×
where the clearance l and pitch
n
S (normal to the flow direction) are illustrated in the sketch on
the next page for tubes in a square pitch.
6
The clearance
n o
l S D = − . When the volumetric flow rate of the shellside fluid is divided by
the crossflow area defined here, it yields the “maximum velocity” through the tube bank,
max
V .
The symbols , , and k ρ µ represent the thermal conductivity, density, and viscosity of the shell
side fluid, respectively, and all the properties should be evaluated at the arithmetic average
temperature of that fluid between the two end temperatures. The symbol Pr stands for the
Prandtl number of the shellside fluid. The exponent n and the multiplicative constant C
depend on the pitch to tube OD ratio, and are given in a table provided in Holman’s book. An
excerpt from the table for tubes on a rectangular pitch (inline tube rows) is given below.
Values of the constant C
/
n o
S D
1.25 1.5 2.0 3.0
/
p o
S D
1.25 0.386 0.305 0.111 0.0703
1.5 0.407 0.278 0.112 0.0753
2.0 0.464 0.332 0.254 0.220
3.0 0.322 0.396 0.415 0.317
clearance
pitch
Flow direction
n
S
l
Tube
OD
o
D
pitch
p
S
7
Values of the constant n
/
n o
S D
1.25 1.5 2.0 3.0
/
p o
S D
1.25 0.592 0.608 0.704 0.752
1.5 0.586 0.620 0.702 0.744
2.0 0.570 0.602 0.632 0.648
3.0 0.601 0.584 0.581 0.608
As an alternative, one can use the procedure outlined in Section 4.5.1 of the book by Mills. For
the shellside heat transfer coefficient, the Nusselt number calculated from correlations using
properties at the arithmetic average of the inlet and exit temperatures is usually sufficient.
The actual flow patterns are more involved, because the flow entering the shell has to distribute
itself into the space in which the tubes are located, and then the flow has to turn around each
baffle. At the exit, the flow again has to converge toward the exit pipe from the shell. In
addition, corrections need to be applied for leakage around the baffles, for bypass of tube
bundles, and other less important nonidealities. As a rough rule of thumb, because of these
various corrections, the ideal heat transfer coefficient
o
h for flow across the tube bank
calculated using a suitable correlation is multiplied by a conservative correction factor of
0.6 in the end.
Pressure Drop
TubeSide Pressure Drop
In designing heat exchangers, pressure drop considerations are usually quite important.
Typically, a design constraint might be P N psi ∆ ≤ , where the number N is specified, and such
constraints may apply on both the tube side and the shell side. Calculation of the tubeside
pressure drop is made by first estimating the (Darcy) friction factor for flow through the tubes
from the value of the Reynolds number and the relative roughness, and applying the viscosity
correction we discussed in class. Then, this friction factor is used to evaluate the pressure drop
for flow through the tubes from
2
1
Number of tube passes
2
corrected
L
P f V
D
ρ
 
∆ = ×

\ .
where L is the length of the tubes, D is the ID of the tubes, ρ is the density of the tubeside
fluid, and V is the average flow velocity through a single tube. To this, we must add
r
P ∆ , the
return pressure loss. This accounts for the pressure drop associated with fluid entry into the tube
bundle, fluid leaving the bundle, and fluid flowing around bends.
8
2
4 Number of tube passes
2
t
r
G
P
ρ
 
∆ = × ×

\ .
Here,
t
G V ρ = is the mass velocity and is defined as
Mass flow rate
Total flow area available per pass
t
t
m
G
A
=
and
Total number of tubes Cross  sectional area of a tube
Number of passes
t
A
×
=
ShellSide Pressure Drop
There are several ways to estimate the pressure drop for the flow of the shellside fluid in a shell
andtube heat exchanger. A reasonable estimate can be obtained by the relatively simple
approach described below, which is given in a book by Peters, Timmerhaus, and West (2). This
book also provides much valuable information on the design of such heat exchangers, including
more sophisticated methods of estimating the pressure drop.
The pressure drop on the shellside is calculated using
( )
2
0.14
2 1
s s B
shell
e
s
f G D N
P
D
µ
ρ
µ
+
∆ =
 

\ .
In this equation, f is a Fanning friction factor for flow on the shell side given in Figure 1444 of
reference (2),
s
G is the mass velocity on the shell side,
s
D is the inside diameter of the shell,
B
N is the number of baffles, ρ is the density of the shellside fluid, and
e
D is an equivalent
diameter. The mass velocity /
s m
G m S = , where mis the mass flow rate of the fluid, and
m
S is
the crossflow area measured close to the central symmetry plane of the shell containing its axis.
This area is defined as
clearance
Cross flow area
pitch
s B
D L = ×
where
B
L is the baffle spacing, and the clearance and pitch are defined in the notes on shelland
tube heat exchangers. The equivalent diameter is defined as follows.
9
2
2 0
0
4
4
p n
e
D
C S
D
D
π
π
 
−

\ .
=
Here,
0
D is the outside diameter of the tubes, and
n
S is the pitch (centertocenter distance) of
the tube assembly. The constant 1
p
C = for a square pitch, and 0.86
p
C = for a triangular pitch.
The friction factor f is given in Figure 1444 of the book as a function of the Reynolds number
based on the equivalent diameter (Note the difference from the Reynolds number that we use for
the heat transfer coefficient from Holman, which uses
0
D as the length scale). For the friction
factor graph, we must use the Reynolds number Re defined as
Re
e s
D G
µ
=
where µ is the viscosity of the shellside fluid. A scanned image of Figure 1444 from Peters et
al. (2) is available for your use at the course web site.
An alternative approach to estimating the shellside pressure drop is given on pages 1110 to 11
11 from Perry’s Handbook; the notation is explained in pages 117 and 118. But, it is
recommended that you use the simple approach given in Peters et al. (2).
Cost
Cost is always an important consideration in designing any process equipment. Cost can be
broken into two principal components – capital cost and operating cost. In addition, maintenance
costs are incurred during operation, but they tend to be more or less independent of the size of
the heat exchanger, so long as the size is within a reasonable range.
The capital cost for heat exchangers increases with increase in the heat transfer area, and is
evaluated by using values known from 19571959, and applying a multiplicative factor known as
the “Cost Index.” This index is published in each issue of Chemical Engineering, and uses 100 a
the basis for the cost in 19571959. To find the cost for the equipment in 195759, consult the
nomogram in Figure 1141 and Tables 1113 and 1114 from Perry’s Handbook.
Operating cost is primarily pumping cost. The pumps must provide work to overcome the
pressure drop on the tube side and that on the shell side. The shaft work per unit mass of fluid is
/ P ρ ∆ , and this must be multiplied by the mass flow rate of the stream to obtain the shaft work
per unit time or shaft power. Then, this must be divided by the overall pump efficiency to obtain
the actual power needed. It is typical to conservatively assume the overall pump efficiency to be
0.6. The yearly pumping power cost can be calculated if one knows the cost per KWH (kilowatt
hour). You can assume operation 24 hours per day for 350 days a year (the remaining days
being nominal maintenance shutdown days).
10
By writing off the capital costs over a certain length of time, the total cost per year can be
worked out. This is then minimized by making a suitable choice of heat exchanger, a job that
requires examining several designs using software to perform the tedious computations.
References
1. J.P. Holman, Heat Transfer, 9
th
Edition, McGrawHill, 2002.
2. Peters, M.S., Timmerhaus, K.D., and West, R.E., Plant Design and Economics for Chemical
Engineers, McGrawHill, New York, 2003.
Heat transfer aspects The starting point of any heat transfer calculation is the overall energy balance and the rate equation. called 11. and the choice is dictated by corrosion concerns as well as mechanical strength requirements.” The sketch schematically illustrates this concept in plan view.hot (Thot . The entire bundle of tubes (typically in the hundreds) is illustrated by a single line in the sketch. because the shell fluid flows across the bank of tubes.The most common material of construction is carbon steel.out − Tcold . Assuming only sensible heat is transferred.out ) Q = mcold C p . Expansion joints are used to accommodate differential thermal expansion of dissimilar materials. Consider the simplest possible shellandtube heat exchanger. which means that there is a single shell “pass” and a single tube “pass. and there are baffles on the shell side to assure that the fluid does not bypass the tube bank.in ) = UA F ∆Tlm Q The various symbols in these equations have their usual meanings. Other materials such as stainless steel or copper are used when needed. The baffle cuts are aligned vertically to permit dirt particles settling out of the shell side fluid to be washed away. we can write the heat duty Q as follows. Note that the contact is not really countercurrent. Baffle T1 t1 t2 T2 2 .cold (Tcold .in − Thot . = mhot C p . The new symbol F stands for a correction factor that must be used with the log mean temperature difference for a countercurrent heat exchanger to accommodate the fact that the flow of the two streams here is more complicated than simple countercurrent or cocurrent flow.
we must find the crossflow velocity across a bundle of tubes that occurs between a pair of baffles. Multiple shell passes also are possible. Here is an example of a 12 (1 shell pass and 2 tube passes) heat exchanger.The convention in shellandtube heat exchangers is as follows: T1 : inlet temperature of the shellside (or hot) fluid T2 : exit temperature of the shellside (or hot) fluid t1 : inlet temperature of the tubeside (or cold) fluid t2 : exit temperature of the tubeside (or cold) fluid Thus. for a variety of situations. Correction factors to be used in the rate equation have been worked out by analysis. the tube outside diameter is employed. but involve fabrication that is more complex and is usually avoided. the formulae for 3 . In the olden days. if possible. Most shellandtube heat exchangers have multiple “passes” to enhance the heat transfer. subject to a set of simplifying assumptions. turns around and flows all the way back. in a 12 heat exchanger. and determine the value of this velocity where the space for the flow of the fluid is the smallest (maximum velocity). For the length scale. Baffle T1 t1 t2 T2 As you can see. It is possible to have more than two tube passes. in evaluating the Reynolds number. For the shell side. the tubeside fluid flows the entire length of the shell. (T1 − t ) − (T2 − t1 ) ∆Tlm = 2 T − t ln 1 2 T2 − t1 The fraction of the circular area that is open in a baffle is identified by a “percentage cut” and we refer to the types of baffles shown as “segmented” baffles.
All elements of a given fluid stream experience the same thermal history as they pass through the heat exchanger (see footnote in Perry for a discussion regarding the violation of this assumption in shellandtube heat exchangers). (or any multiple of 4) tube passes. 1. given in Perry’s Handbook. are as follows. respectively. Figures C4. Nowadays. 2. 4.them were considered too cumbersome to use. (or any multiple of 2) tube passes. A misleading rule of thumb is that F ≥ 0.aT1 − t1 t2 − t1 d in Appendix C of the textbook by Mills display such graphs. The formula given above for F1− 2 also applies for one shell pass and 2. 8. The overall heat transfer coefficient U is constant. P B= 2 P F1− 2 F2− 4 where A = (1 − P )(1 − PR ) The first and second subscripts on the factor F correspond to the number of shell and tube passes. R2 + 1 1 − P ln R − 1 1 − PR = A + R2 + 1 ln 2 A − R +1 R2 + 1 1 − P ln 2 ( R − 1) 1 − PR = A + B + R2 + 1 ln 2 A + B − R +1 2 −1− R . 3.8 . Given next are the two common factors. The heat exchanger is at steady state. 5. In designing heat exchangers. but the correct idea is that the region of steep falloff in the curves should be avoided. one should avoid the steep portion of the curves of F versus P . the formula for F2− 4 also applies for two shell passes and 4. 4. Therefore graphs were prepared plotting t −t T −T F ( P. R ) . Heat losses are negligible. where P = 2 1 and R = 1 2 are parameters on which F depends. 4 . one can compute these factors quickly with a pocket calculator. because small errors in estimating P can cause large changes in the value of F . The simplifying assumptions mentioned in the previous paragraph. Likewise. The specific heat of each stream remains constant throughout the exchanger.
Heat transfer correlations for flow through tube banks are used.i U o ho k Alm hi Ai In the above equation. where Do is the outside diameter of a tube. such as those given in the book by Holman (1). The correlation given in Holman’s book is = Nu ho Do = C Re n Pr1/ 3 k The Reynolds number Re = DoVmax ρ µ . The inside heat transfer coefficient hi can be evaluated using the standard approach for predicting heat transfer in flow through tubes. You’ll find several tables of typical overall heat transfer coefficients in shellandtube heat exchangers in Chapter 11 of Perry’s Handbook.0 for the shell side.Heat Transfer Coefficients The evaluation of the overall heat transfer coefficient is an important part of the thermal design and analysis of a heat exchanger. These correlations assume flow normal to the long axes of a set of tubes placed in a geometrical array. The following generic result can be written for the overall heat transfer coefficient U o based on the outside surface area of the tubes. Predicting the shellside heat transfer coefficient ho is more involved. and a good design would aim to arrange for turbulent flow. including applying a viscosity correction where possible. This is given as Cross flow area =ID × Baffle spacing × Shell pitch where the clearance l and pitch S n (normal to the flow direction) are illustrated in the sketch on the next page for tubes in a square pitch.0 + R f . 5 . Vmax is the “maximum” velocity of the fluid through the tube bank. and R f . Typically. and Alm is their log mean. the crossflow area clearance must be evaluated. The presence of baffles needs to be taken into account in calculating the fluid velocity across the tube bank. respectively. which is the heat transfer surface. even in the absence of baffles. first. Ai and Ao are the inside and outside surface areas of a tube. turbulent flow can be expected. Accumulated information on fouling resistances can be found in the Standards published by TEMA. hi is the heat transfer coefficient for the fluid flowing through the tubes. because of the substantial enhancement in heat transfer provided by eddy transport. because the flow passage is not simple. To find it.i for the tube side. The fouling resistances on a unit area basis are R f . 1 1 ∆ r Ao 1 Ao = + + + R f . ho is the heat transfer coefficient for the fluid flowing in the shell.
25 0.396 2.5 0. Values of the constant C S n / Do S p / Do 1.0 6 .0 0. When the volumetric flow rate of the shellside fluid is divided by l the crossflow area defined here.317 1. density. and viscosity of the shellside fluid.305 0. An excerpt from the table for tubes on a rectangular pitch (inline tube rows) is given below.0 3. respectively.111 0. The symbols k . and all the properties should be evaluated at the arithmetic average temperature of that fluid between the two end temperatures.407 0.Flow direction clearance l pitch Do Tube OD Sn pitch S p The clearance = S n − Do . and µ represent the thermal conductivity. and are given in a table provided in Holman’s book.112 0.322 1. ρ .0703 0.25 1. it yields the “maximum velocity” through the tube bank.386 0.5 2. The symbol Pr stands for the Prandtl number of the shellside fluid.0753 0.0 0.278 0.332 0.254 0.464 0. The exponent n and the multiplicative constant C depend on the pitch to tube OD ratio. Vmax .415 3.220 0.
In addition. Typically.25 0.25 1.608 1.586 0. Calculation of the tubeside pressure drop is made by first estimating the (Darcy) friction factor for flow through the tubes from the value of the Reynolds number and the relative roughness.581 3. ρ is the density of the tubeside fluid. For the shellside heat transfer coefficient.0 0.1 of the book by Mills. To this. At the exit.744 0. the return pressure loss. a design constraint might be ∆P ≤ N psi .0 As an alternative.5 0. The actual flow patterns are more involved.584 2.620 0. the ideal heat transfer coefficient ho for flow across the tube bank calculated using a suitable correlation is multiplied by a conservative correction factor of 0.Values of the constant n S n / Do S p / Do 1. As a rough rule of thumb.0 0. the flow again has to converge toward the exit pipe from the shell. Then. because the flow entering the shell has to distribute itself into the space in which the tubes are located. and other less important nonidealities. for bypass of tube bundles. where the number N is specified.632 0.5 2.6 in the end.752 0. pressure drop considerations are usually quite important.601 1. because of these various corrections.648 0.704 0. 7 . and then the flow has to turn around each baffle. D is the ID of the tubes. and applying the viscosity correction we discussed in class.5.702 0.0 3. we must add ∆Pr . fluid leaving the bundle. and V is the average flow velocity through a single tube. and fluid flowing around bends.608 0. one can use the procedure outlined in Section 4. Pressure Drop TubeSide Pressure Drop In designing heat exchangers.602 0. and such constraints may apply on both the tube side and the shell side.592 0.570 0. corrections need to be applied for leakage around the baffles. This accounts for the pressure drop associated with fluid entry into the tube bundle. the Nusselt number calculated from correlations using properties at the arithmetic average of the inlet and exit temperatures is usually sufficient. this friction factor is used to evaluate the pressure drop for flow through the tubes from = ∆P f corrected L 1 2 ρV × Number of tube passes D 2 where L is the length of the tubes.
f is a Fanning friction factor for flow on the shell side given in Figure 1444 of reference (2). This area is defined as Cross flow = Ds LB × area clearance pitch where LB is the baffle spacing. and West (2). The equivalent diameter is defined as follows. Gt = ρ V is the mass velocity and is defined as Gt = Mass flow rate m Total flow area available per pass At Total number of tubes × Cross . which is given in a book by Peters. and De is an equivalent diameter. ρ is the density of the shellside fluid. including more sophisticated methods of estimating the pressure drop. The pressure drop on the shellside is calculated using ∆Pshell 2 f Gs2 Ds ( N B + 1) = 0. A reasonable estimate can be obtained by the relatively simple approach described below. G2 ∆Pr = 4 × Number of tube passes × t 2ρ Here. where m is the mass flow rate of the fluid.14 µ ρ De µs In this equation. This book also provides much valuable information on the design of such heat exchangers. Ds is the inside diameter of the shell. Gs is the mass velocity on the shell side. and the clearance and pitch are defined in the notes on shellandtube heat exchangers. 8 . Timmerhaus. N B is the number of baffles. and S m is the crossflow area measured close to the central symmetry plane of the shell containing its axis. The mass velocity Gs = m / S m .sectional area of a tube Number of passes and At = ShellSide Pressure Drop There are several ways to estimate the pressure drop for the flow of the shellside fluid in a shellandtube heat exchanger.
and applying a multiplicative factor known as the “Cost Index. In addition.6.86 for a triangular pitch. which uses D0 as the length scale). so long as the size is within a reasonable range. For the friction factor graph. Cost Cost is always an important consideration in designing any process equipment. and S n is the pitch (centertocenter distance) of the tube assembly. An alternative approach to estimating the shellside pressure drop is given on pages 1110 to 1111 from Perry’s Handbook. A scanned image of Figure 1444 from Peters et al. but they tend to be more or less independent of the size of the heat exchanger. It is typical to conservatively assume the overall pump efficiency to be 0. it is recommended that you use the simple approach given in Peters et al. The shaft work per unit mass of fluid is ∆P / ρ . Cost can be broken into two principal components – capital cost and operating cost. The pumps must provide work to overcome the pressure drop on the tube side and that on the shell side. and is evaluated by using values known from 19571959. consult the nomogram in Figure 1141 and Tables 1113 and 1114 from Perry’s Handbook. the notation is explained in pages 117 and 118. To find the cost for the equipment in 195759. maintenance costs are incurred during operation.” This index is published in each issue of Chemical Engineering. and C p = 0. we must use the Reynolds number Re defined as Re = DeGs µ where µ is the viscosity of the shellside fluid. 9 . But. and this must be multiplied by the mass flow rate of the stream to obtain the shaft work per unit time or shaft power. The constant C p = 1 for a square pitch. (2) is available for your use at the course web site. The friction factor f is given in Figure 1444 of the book as a function of the Reynolds number based on the equivalent diameter (Note the difference from the Reynolds number that we use for the heat transfer coefficient from Holman. Operating cost is primarily pumping cost. π D02 2 4 C p Sn − 4 De = π D0 Here. (2). You can assume operation 24 hours per day for 350 days a year (the remaining days being nominal maintenance shutdown days). The capital cost for heat exchangers increases with increase in the heat transfer area. D0 is the outside diameter of the tubes. Then. this must be divided by the overall pump efficiency to obtain the actual power needed. The yearly pumping power cost can be calculated if one knows the cost per KWH (kilowatthour). and uses 100 a the basis for the cost in 19571959.
This is then minimized by making a suitable choice of heat exchanger..P. R. 2003.. 9th Edition.E. Heat Transfer. M. References 1. 2. J. and West. K. Peters. Holman. a job that requires examining several designs using software to perform the tedious computations.By writing off the capital costs over a certain length of time. the total cost per year can be worked out. Plant Design and Economics for Chemical Engineers.D. Timmerhaus. 2002. New York. McGrawHill. McGrawHill.S.. 10 .
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.