6 CONVECTION HEAT TRANSFER IN CLOSED

CONDUIT
The most important convective heat-transfer process industrially is that of cooling or
heating a fluid flowing inside a closed conduit or pipe. The relationship between
temperature profile and heat transfer at a bounding surface is defined in terms of the
heat transfer coefficient introduced in Newton’s law of cooling. The heat transfer
coefficient depends on the temperature difference selected. For flow in a conduit, there
are several temperature differences of importance, and each has an associated heat
transfer coefficient.

Although the solution for the differential energy balance in the laminar flow regime is
available, engineers prefer to work with heat transfer coefficients. As a result, it is
customary to express heat transfer performance, even for laminar flow, in terms of
heat transfer coefficient.

The heat transfer coefficients for laminar flow have a strong dependence on position.
This is not usually the case for heat transfer with turbulent flow.

There are 2 objectives:

(1) To relate the heat transferred to the temperature change experienced by the fluid,
(2) To relate the heat transferred to the fluid to the difference in temperature between
the wall and the fluid

6.1 Laminar flow, entrance region

Consider a reservoir from which fluid flows under laminar conditions into a circular
duct. The duct entrance is well rounded. The mouth of the tube is located at z = 0.

Uniform entrance Developing velocity

velocity profile vz (r,z)

vmax = 2vavg

vz(r)

z
Fully developed region

(developing flow) ze
Development of the hydrodynamic boundary layer in a pipe

83
A fluid usually enters a pipe with uniform velocity (vavg) and temperature (T∞)
profiles. If the pipe wall is heated or cooled, then both thermal and hydrodynamic
boundary layers begin developing at the pipe inlet. The boundary layers coat the inner
surface of the tube. As the boundary layers grow in thickness, eventually they fill the
tube. Beyond this location, the velocity profile or the temperature distribution no
longer changes, the flow becomes hydrodynamically and/or thermally fully developed.

As the fluid moves along the pipe, the velocity profile changes from being a constant
at inlet (vavg) to fully developed downstream. A fully developed velocity profile does
not change any longer in the flow direction. The axial distance required for the
velocity profile to become fully developed is called the hydrodynamic entrance
length ze. At this distance the boundary layer fills the pipe cross-section, that is δ = R
at z = ze.

TW > T∞
q”W
TW > T∞

T∞

NO ARROWS!
T∞ T∞ TW T∞ TW T∞ TW

z zT

Development of thermal boundary layer in pipe and temperature profiles for

(B) constant TW, (C) constant q″

When the fluid enters the pipe, convective heat transfer will occur, a thermal boundary
layer will start to develop, and the fluid’s uniform T∞ temperature will increase. In the
thermal entrance region (A), the temperature of the central portion of the flow outside
the thermal boundary layer δT remains unchanged, but in the boundary layer the
temperature increases sharply to the pipe surface temperature TW. The thermal
boundary layer builds up till it fills the pipe cross-section. After this point, fully
developed flow condition has been reached.

In the fully developed region, the fluid at the centerline begins to heat up as well.

The thermal entry length zT is where δT = R.

In laminar flow: ze ≅ 0.05D Re where

ze = hydrodynamic entrance length
zT ≅ 0.05D Re Pr zT = thermal entry length

84
In turbulent flow: ze ≅ 4.4D Re1/6 and zT ≅ 10D

The entry length in turbulent flow is much shorter as compared to laminar flow. In
most engineering applications, the entrance effect becomes insignificant beyond the
pipe length of about 10D.

If the total pipe length is large compared to the entrance length, then the effect of the
entrance length can be neglected and the total pipe length can be used in calculations.

If the total pipe length is relatively short in comparison with the entrance length,
however, then the entrance region needs to be analysed separately.

For gases (Pr < 1), therefore, the hydrodynamic boundary layer develops more slowly
than the thermal boundary layer (ze > zT). For liquids (Pr > 1), the flow develops
hydrodynamically more rapidly than it does thermally. For substances with high
Prandtl numbers, such as oils (Pr > 100), the flow develops hydrodynamically so much
more quickly that a fully developed velocity profile can be reasonably assumed
throughout.

δ δT
δT δ
Pr >> 1 Pr << 1
(e.g. oils) (e.g. liquid metals)

A fully developed velocity profile is mathematically defined as:

d (v z / v z max )
= 0 ⇒ v z = v z (r ) it is independent of z.
dz
A fully developed temperature profile is one in which

d [ (TW − T ) /(TW − Tb ) ] where TW = wall temperature

= 0
dz Tb = bulk fluid temperature or
mean fluid temperature

These equations are valid at each radial location in the cross section.

Unlike the velocity profile, the temperature profile can be different at different cross
sections of the tube in the developed region, and it usually is. However, above
temperature ratio (dimensionless temperature profile) remains unchanged when TW or
q″ is constant.

85
 TW = constant
x2 > x1

Tm ≡ Tb
bulk or mean
temperature

Temperature profiles for thermally fully developed flow for constant temperature
heating. The relative shape of the temperature profiles is the same.

The bulk temperature for a circular conduit (ρ, cp = constant) is:

2π R

∫ ∫ v z (r ) T (r , z ) r dr dΘ
Tb ( z ) =
(v z T ) avg
=
∫∫ v z T dA = 0 0
2π R
v z ,avg ∫∫ v z dA
∫ ∫ v z (r ) r dr dΘ
0 0

The local heat transfer coefficient h z is: q = hz AS (TW – Tb)

TW – T b is the temperature difference between the wall and the fluid. It is an

important quantity. Both TW and Tb can vary in the z-direction.

If we integrate the local coefficient hz over a definite area (such as the surface area
AS), the integrated result would be h average heat transfer coefficient:
1
h=
As ∫ hz dAs
AS

The solution in the entrance region

Nusselt number, thus h heat transfer coefficient is much higher in the entrance region.

The Leveque solution (1928) for the Nusselt number in the entrance region for laminar
flow and constant wall temperature
1 /3
D D
Nu = 1.077 (Re Pr) 1 /3
  for 100 < Re Pr < 5000
z z

86
6.2 Heat transfer in laminar flow — Constant heat flux at
the wall
The heat flux at the wall q″w = q/AS = constant. Experimentally this can be achieved
by passing electric current through a metal pipe or by wrapping the duct with a
material through which an electric current is passed. The resulting temperature-profile
variation with axial distance is illustrated in the next figure.

The velocity profile and the temperature profile both begin changing at the tube
entrance. The fluid temperature at the inlet is T∞, the tank-fluid temperature and varies
within the tube. At any axial location z, T varies from TW, the wall temperature, to
some value at the centerline. Just as the velocity profile eventually becomes fully
developed, so does the temperature profile. The temperature profile reaches a non-
changing shape, but with the constant heat input that exists, T−T∞ increases with
increasing axial distance z.

TW > T∞

~~~~
TT∞
4
TW TW \ q”w = q /A s = constant
TW TW

T∞
T∞4 T∞4 T T∞4
T T
T∞4
4
T (z,r)

tank TW (z)

T hz T b (z)
hz
q/As ΔT = TW – Tb

q”w = constant

thermally developed region

0 zT z

The figure shows also the variation of the wall temperature and the bulk temperature
with increasing z. The bulk temperature increases linearly with distance z. The wall
temperature changes over the thermal entry length zT. Beyond zT, the wall- and bulk-
temperature lines are parallel.

87
The temperature difference between the wall and the fluid (TW – Tb) is used in
Newton’s law of cooling to define a local heat transfer coefficient:

q″ = hz (TW – Tb)

In the entrance region: TW and Tb vary with z ⇒ hz also varies with z

In the thermally developed region: TW – Tb = constant ⇒ hz = constant
q′′
and TW − Tb = ∆T =
hz
Note: hz is infinite at the entrance of the tube.

The relationship between Tb and q″

q” w = constant

r Tb Tb + dTb P = 2πR = perimeter of tube

z A = πR2 = flow cross-section

dz
q”w
Perform a heat balance for the element:
dq dq
q ′w′ = = ⇒ dq = q ′w′ P dz
dAS P dz

The heat added increases the enthalpy of the fluid:

=

dq = m dh = m c p dTb = ρ vavg A c p dTb

2π R q′w′ dz = ρ vavg π R 2c p dTb
⇓
dTb 2 q ′w′ α k
= = constant where α =
dz vavg k R ρ cp

The slope of the bulk temperature line T b(z) is constant ⇒ Tb varies linearly with z,
this result is independent whether or not fully developed conditions exist.

Integrating from Tbi to Tbo and from 0 to L yields:

2 q′′α L
Tbo − Tbi = applies to laminar or turbulent flow regime
vavg k R

The bulk temperature Tb

11 q′w′ D
TW − Tb =
48 k

88
The heat transfer coefficient
Introducing above result into Newton’s law of cooling [q″w= hz (TW – Tb) ] results

hz D 48 constant wall heat flux

Nu = = = 4.364 hydrodynamically and thermally developed
k 11 laminar flow

6.3 Heat transfer in laminar flow — Constant wall

temperature
A fluid at temperature T∞ flows from a reservoir through a pipe of constant wall
temperature TW (>T∞). The resulting variation of temperature profile with axial
distance is illustrated in the figure below.

TW > T ∞4
~~~~ TW TW TW TW
TT∞4

T∞4 T ∞4 T ∞4 T ∞4 T∞4
T (z,r)

tank

T
TW
hz
q/As
ΔT = TW – Tb T b (z)

hz

q”w
T4
zT z
0

Experimentally constant wall temperature can be achieved by allowing a fluid to

condense on the outside surface of the tube. Condensation is an isothermal process for
pure substances.

− Tb changes from inlet and approaches the constant TW with increasing z.

− Tb does not increase linearly.
− q″ heat flux decreases with increasing z.
− ΔT = (TW – Tb) difference decreases with z, but so does the heat flux q″. The
combined effect is that the local heat transfer coefficient hz becomes constant in
the thermally developed region. [ q″w = hz (TW – Tb) ]
− ΔT decays exponentially.
89
The heat transfer rate

(TW − Tbo ) − (TW − Tbi )

q = 2 π R L hL
T − Tbo
ln W
TW − Tbi

2πRL = AS the surface area of a circular duct

Introducing the logarithmic temperature difference ΔTlm:

(TW − Tbo ) − (TW − Tbi )

∆Tlm ≡
T − Tbo
ln W
TW − Tbi

q = hL AS ∆Tlm for laminar flow, TW = constant

Note:
− the integrated coefficient hL is defined in terms of ΔTlm
− the local coefficient h z is defined in terms of (TW − Tb)
− ΔTlm is an exact representation of the changing average temperature difference
between the fluid and the surface.
− ΔTlm should always be used when determining the convection heat
transfer in a tube whose surface is maintained at a constant temperature
TW.

Above equations and results do not require that fully developed conditions exist.

The heat transfer coefficient

constant wall temperature

h D
Nu = z = 3.658 hydrodynamically and thermally developed
k laminar flow

Graphical representation of results

The Nusselt number based on a local coefficient hz is a significant parameter. From
dimensional analysis it can be seen that

hz = hz (vavg , ρ, μ, D, c p , k, z)

90
 hz D z z
⇒ = h z (Re, Pr, ) or Nu = Nu (Re, Pr, )
k D D

Defining Grätz (Graetz) number as:

D
Gz ≡ Re Pr so Nu = Nu(Gz) or Nu = Nu(1/Gz)
z
The following graph summarizes the results obtained from the solution of the
differential equation of energy. The temperature profile is thermally developed when
1/Gz ≥ 0.05.

20
Nu (1) (1) avg. Nu at TW = const.

(2)
(2) local Nu at q”W = const.
10
(3) local Nu at T W = const.
(3) 48 / 11
3.658
0
0.001 0.01 0.1 1.0

1/Gz = z / (D Re Pr)

Local and average Nusselt numbers for circular tube thermal entrance regions in fully
developed laminar flow.

91
6.4 The combined-entry-length problem for laminar flow
When the laminar flow is not developed hydrodynamically or thermally, the case is
referred to as the combined-entry-length problem. For this case, the heat transfer in the
entrance region is more sensitive to the Prandtl number.

Graphical representation of the problem (as graphs of Nu vs. 1/Gz) is available in

literature for constant wall temperature and for constant wall heat flux as well. It is
often impractical to have graphs from which to obtain Nusselt number.

For the combined problem of developing velocity and temperature profiles, the
Sieder-Tate equation is useful:
1 /3 0.14
h D  D  µ 
Nu = L = 1.86  Re Pr   
k  L  µw 

where μw = dynamic viscosity of fluid at TW

h L = average heat-transfer coefficient over the heat-transfer length
Fluid properties (k, ρ, c p , μ) are evaluated at the average of the bulk temperature:

Tbo + Tbi
Tavg =
2

It applies to steady laminar flow of Newtonian fluid in a tube having constant wall
temperature under the following conditions:

TW = constant (in condensers when heating a fluid with condensing vapor)

0.48 < Pr < 16 700 ( for gases, organic liquids, water solutions −but not water)
0.044 < (μ /μ w) < 9.75
Re < 2100 (laminar flow)
Re Pr (D/L) > 10
Deviation: ± 12%

In laminar flow, the average coefficient h L depends strongly on the heated length.
The average temperature drop (ΔTavg) between the bulk fluid and the wall is used to
calculate the heat-transfer rate:

(TW − Tbi ) + (TW − Tbo )

q = h A ∆Tavg = h A
2

92
6.5 Heat transfer in turbulent flow in tubes
In turbulent flow, the hydrodynamic and thermal entry lengths are of the same size and
are independent of the Prandtl number. The entry length is much shorter than in
laminar flow and its dependence of the Reynold number is weak. The entrance effect
becomes insignificant beyond a tube length of 10 diameters:

z e ≈ zT ≤ 10 D

The tubes used in practice in forced convection are usually several times longer than
the length of the entrance region, and thus we assume fully developed flow for the
entire length of the tube.

Heat transfer coefficients are higher with turbulent flow than with laminar flow.

The Nusselt numbers for TW = constant or q″w = constant are identical in the fully
developed regions, and nearly identical in the entrance regions.

It is not possible to obtain a closed-form solution for the velocity profile in turbulent
flow. Dimensional analysis predicts that the Nusselt number depends on Reynolds and
Prandtl numbers. Combining with experimental results, the following is found to
apply:

Sieder-Tate equation (1936)

0.14
h D  µ 
Nu = L = 0.027 Re 0.8 Pr1 /3  
k µ
 w

Conditions:
TW = constant or q″w = constant
Re > 10 000 (turbulent flow)
0.7 ≤ Pr ≤ 16 700
L /D ≥ 60 thermally and hydrodynamically developed flow
Deviation: +15% to −10%

hL is the heat-transfer coefficient based on the log mean driving force ΔTlm:

q = hL A ΔTlm

This correlation is applicable for fluids whose viscosity changes greatly for small
temperature change. It is not recommended for water.

The use of this equation may be trial and error, since value of hL must be known to
evaluate TW, and hence μ w , at the wall temperature.

Since the rate of heat transfer is greater in turbulent flow than in laminar flow, most
equipment are operated in the turbulent range.

93
In place of this type of correlation, another is often used, which consists of using a
dimensionless group j called the Colburn factor (1933)

where G = ρv = mass velocity or mass flux,

 h   c p µ  2 /3 kg/m2s
j =   
 c pG   k  h/cpG = St = Stanton number
  cpμ/k = Pr = Prandtl number

⇒ j = St Pr2/3
f =
Colburn showed that j = F
2

By this, Colburn related the Fanning friction factor to the heat transfer coefficient.
The dimensionless group j is often plotted against the Re number.

Dittus-Boelter equation (1930)

It is an alternative expression that is useful when the wall temperature is unknown.

Nu = 0.023 Re0.8 Prn

where n = 0.4 for heating ( TW > Tb )

n = 0.3 for cooling ( TW < Tb )

Conditions:
TW = constant or q″w = constant
0.7 ≤ Pr ≤160
Re > 10 000 (maybe Re > 8 000) turbulent flow
L / D ≥ 60
Deviation: ± 20%

This correlation is applicable in case of moderate temperature differences between the

wall and the fluid. This equation should be used for water.

The transition region extending from Re = 2100 to 10000 is not well understood and is
usually avoided in design if possible.

These correlations for the turbulent regime predict the local Nusselt number. The
average Nusselt number based on a heat transfer coefficient for the entire tube is not
available.

When dealing with non-circular conduits, the hydraulic diameter or equivalent

diameter is used in the heat transfer correlations whenever D diameter appears.

cross sectional area

De = 4 ⋅ rH = 4
wetted perimeter
94
 Summary of heat transfer correlations in forced convection
CORRELATION EQUATION RESTRICTIONS
Fully developed turbulent
flow in pipes
Dittus – Boelter eq. Nu = 0.023 Re0.8Prn n = 0.4 for heating
(1930) n = 0.3 for cooling
TW = const. or q"w = const.
2500 < Re < 105
0.7 ≤ Pr ≤160
Fully developed turbulent
0.14
 μ  flow in pipes
Sieder – Tate eq. Nu = 0.027Re0.8Pr1/3  μ 
TW = const. or q"w = const.
 W 
(1936) Re > 10 000
0.7 < Pr < 17 600
1 0.14
 d 3  μ  Fully developed laminar
Sieder – Tate eq. Nu = 1.86 (RePr)1/3  L    flow in pipes
 μW  TW = const.
RePr(d/L) > 10
f 
Second Petukov eq.   RePr For smooth pipe
(1970) Nu = 8 104 < Re < 5x106
( )
0.5
f  0.5 ≤ Pr ≤ 2000
1.07 + 12.7   Pr 2/3 − 1
8
f 
  (Re − 1000 )Pr
Nu =  
8
Gnielinsky eq. 3000 < Re < 5x106
( )
0.8
f 
(1976) 1 + 12.7   Pr 2/3 − 1 0.5 < Pr < 2000
8

where f = (0.790 lnRe – 1.64) – 2

Approximate magnitude of some heat-transfer coefficients

Mechanism h, W/m2K
Condensing steam 5500-30 000
Condensing organics 1000-3000
Boiling liquids 1500-30 000
Moving water 250-17 000
Moving hydrocarbons 50-1700
Still air 5-25
Moving air 10-55

95
7 HEAT EXCHANGERS

A heat exchanger is a device designed for exchanging heat between 2 fluids.

7.1 Classification of heat exchangers

Heat exchangers can be classified in a number of ways, depending on their
construction or how the fluid move relative to each other through the device.

(1) By flow arrangement:

− parallel flow — the two fluid streams travel in the same direction along the
exchanger,
− counterflow (or countercurrent flow) — the two fluid streams travel in the
opposite direction,
− crossflow — the two fluid paths cross each other at right angles.

(2) By construction:
− double-pipe heat exchanger — It consists of two concentric pipes. One fluid
flows through the inner pipe, the other flows
through the annulus.
− shell-and-tube heat exchanger —It consists of a huge outer cylinder (shell) within
which are contained many tubes (tube bundle).
One fluid flows through the shell, the other
flows through the tubes.
− others (pipecoils, plate-type heat exchanger, extended surface heat exchangers)

(3) By the number of passes: – single pass

– multi pass

The objective is to predict:

− the amount of heat transferred,
− the outlet temperature of one or both fluid streams,
− the required heat transfer area, depending on what is known, and
− overall heat transfer coefficient.

It is usually assumed, that all the heat lost by the hot fluid is transferred to the cold
fluid (that is there are no heat losses).

96
7.2 Temperature distribution
PARALLEL FLOW COUNTERFLOW
t1 t2

T1 T2 T1 T2

t2 t1
T T
T1 T1

T2 )∆T
T11
)∆T
T11 )∆T
T == TT!– tt )∆T2
t2 t2 )∆T
T = T !– t
T2
∆
) T2
t1 t1

Length Length
or area or area

ΔT1 = T1 − t1 ΔT1 = T1 − t2
ΔT2 = T2 − t2 ΔT2 = T2 − t1

where ΔT1 and ΔT2 = terminal temperature differences

T = hot fluid
t = cold fluid
1 = inlet
2 = outlet

In counterflow arrangement t2 (cold fluid outlet temperature) can be greater than T2

(hot fluid outlet temperature). In parallel flow, however, the upper limit of t2 is T2 .

7.3 Double-pipe heat exchanger

One fluid flows through the center pipe while another flows in the annulus. Such an
exchanger may consist of several passes arranged in a vertical stack. The two lengths
of inner pipe are connected by a return bend, which is usually exposed and does not
provide effective heat transfer surface. Two legs make a hairpin.

Usually common copper water tubing is used to construct the exchanger, although pipe
and special tubing can also be used.

97
Effective lengths: – 12 ft (3.65 m)
– 15 ft (4.57 m)
– 20 ft (6.1 m)

Standard sizes of double-pipe exchangers

Outer pipe Inner pipe

2 in (50.8 mm) 1¼ in (31.75 mm)
2½ in (63.5 mm) 1¼ in (31.75 mm)
3 in (76.2 mm) 2 in (50.8 mm)
4 in (101.6 mm) 3 in (76.2 mm)

98
Application of double-pipe exchangers:
– It is useful, when not more, than 9-14 m2 of heat transfer surface is required.
– It is preferred for small capacity, high pressure, and countercurrent operation.
– For low flow rates,
– If the heat duty (q) is less than 500 kW,
– If for the shell-and-tube heat exchanger the number of tubes are less than about 30
(¾ in = 19 mm OD tubes) and the diameter of the shell is less than 200 mm.

Under these conditions, a shell-and-tube exchanger is uneconomic; a large number of

small diameter shells in series is required to provide adequate velocities and near-
countercurrent flow.

Disadvantages:
– Each double-pipe exchanger introduces no fewer than 14 points at which leakage
might occur.
– Considerable time and expense required for dismantling and periodically cleaning it
compared with other types of equipment.
– When a large number are connected, they require considerable space.

99
7.4 Shell-and-tube heat exchanger
Where a high flow rate is involved, the number of double-pipe heat exchangers
required becomes large (high floor space requirement, high capital cost), so an
alternative apparatus, a shell-and-tube heat exchanger can be used. Shell-and-tube heat
exchangers become the industry standard because of several advantages:
− high flow rates
− compactness (large ratio of heat transfer area to volume and weight)
− good design methods exist
− ease of construction in a wide range of sizes
− ease of maintenance
− high heat transfer coefficients
− can handle either high or low pressures and temperatures

The Tubular Exchanger Manufacturer’s Association (TEMA), as an association of heat

exchanger equipment fabricators, has developed a standard for basic construction.

Shell-and-tube heat exchanger with 1 tube pass

The number of tubes in the

Tube Shell tube bundle ranges from about
intlet intlet
20 to over 1000. The tubes are
Nozzle Shell Baffle Tube sheet attached to perforated flat
plates, termed tube sheets, at
each end. The tubes pass
through a number of flat
plates, called baffles, along
their length, which serve to
support them and direct the
fluid flow in the shell in such
Tubes a way that heat transfer is
enhanced. The assembly of
Shell Tube
Single pass 1 – 1 exchanger outlet
tubes and baffles, called the
outlet
100
tube bundle, is held together by a system of tie rods, and spacer tubes.

t2 T1
The simplest multi-pass
shell-and-tube exchanger is
the 1-2 type. This
configuration has higher
t ci
velocities and shorter tubes.
The higher velocity increases
the heat transfer coefficient.
t1 T2
1 – 2 exchanger It has an important limitation:
T1 because of the parallel flow
pass, the exchanger is unable
t2 to bring the exit temperature
of the cold fluid very near to
T2 the entrance temperature of
T t ci the hot fluid.

Further increasing the

number of passes (e.g. 1-4, 1-
t1
6, 2-4, 2-6, 2-8 etc.), the fluid
length velocities could be further
or area
increased.

Shell
The type of shell often is established before thermal rating of the unit takes place.

The TEMA standard type of designated system:

E F G
One pass shell Two pass shell with Split flow
longitudinal baffle

H J X
Double split flow Divided flow Cross flow
Shell types

101
Shell inside diameters are standardized. Up
to 23 in (60 cm) diameters are fixed. Sizes
above 60 cm are specified to the nearest inch
not exceeding 60 in (152.4 cm). Shells are
constructed of rolled plates. Pressures are
not exceeding 200 bar. A B
Channel and Bonnet
removable cover (integral cover)

Front and stationary head types

L Fixed tube sheet like “A”

stationary head
M Fixed tube sheet like “B”
stationary head
P Outside packed floating head
S Floating head with backing
device
T Pull through floating head
Floating head U-tube bundle U U-tube bundle
Rear end head types W Externally sealed floating
tube sheet

Advantages of fixed tube sheet exchanger:

− There are no internal joints, thus eliminating a potential source of leakage of one
fluid into another
− The peripheral tubes may be placed close to the inside of the shell, due to the
absence of an internal joint, so that more tubes can be accommodated in a shell
of given diameter for the fixed tube sheet type than for any other type.
− In addition to the fact that there are no flanges on the shell side makes it desirable
for high-pressure and/or lethal services.
− U-tube has no internal joints either, but there is a practical limit to the bend of the
internal tubes; the number of tubes accommodated within a given shell diameter
is slightly less, than that for the fixed tube sheet type.

Fixed tube sheet may be used with maximum 120°C temperature difference between
the 2 fluids; floating head may be used above 180°C temperature difference.

Shell side velocities: 0.6 – 1.5 m/s for water and water-like materials.

102
Baffles
Baffles are very important part of the performance of
% cut based on diameter a heat exchanger.

Shell side baffles

Shell-side baffles have their thermal significance in
Baffle Baffle that the shell-side fluid is made to flow to and fro
window across the bundle from one end of the exchanger to
the other. There are only a few popular and practical
arrangements.

Removed from baffle The cut-out portion represents the free flow area for
to allow for liquid drainage. the shell side fluid. It is usually 25% of the inside
diameter of the shell.

Baffles can be arranged to direct the shell side fluid in “side-to-side” or “up-and-
down” flow.

This type of baffle is used for:

− single phase clean operation,
− vaporizing, and
− single phase fouling operation,
− condensing.
Standard segmental baffle
designed for side to side flow (vertival cut)

This type of baffle is used for:

− single phase clean operation
(horizontal cut is more
common), and
− vaporizing.
Standard segmental baffle
designed for up and down flow (horizontal cut)

The baffles serve as tube support as well. The baffles are supported by 1 or more guide
rods, which are fastened between the tube sheets.

Baffles in a small shell-and-tube

heat exchanger, only 6 of the 37
tubes are in position.

103
Baffle thickness: it depends on the shell diameter and unsupported tube length, but it
is usually between 1/8 - 3/4 in (3.2-19 mm).
Baffle clearance: the holes for the tubes are drilled 0.4-1 mm larger than the tube Do
outside diameter. The clearance between shell and baffles are usually 3-12 mm.
Baffle spacing: The baffle spacing varies between 1/5 - 1 times the inside diameter of
the shell. The distance between baffles should not be less than 1/5 the diameter of the
shell or 2 in (50.8 mm), whichever is the greater.

Tube side baffles

They are built into the head and return ends of an exchanger to direct the fluid through
the tubes at the proper relative position in the bundle for good heat transfer. The
arrangement may take any of several designs depending on the number of tube side
passes required. The number of tubes per pass is usually arranged about equal.

Tubes
There are 2 basic types of tubes: − plain
− finned (external or internal)

Plane tubes are used in the usual heat exchange applications. Plain tube mechanical
data and dimensions are given in tables. The usual tube sizes for most exchangers are
¾ in and 1 in OD (19.05 and 25.4 mm). Tubes of 1¼ in and 1½ in OD and sometimes
larger are used in boilers, evaporators, and reboilers. They are available in all common
metals. Standard lengths of tubes for heat exchanger construction are 8, 12, 16, 20, and
24 ft (2.438, 3.658, 4.877, 6.096, and 7.325 m) for both straight and U-tubes, but other
lengths might be used.

Tube side velocities: 0.9 – 2.4 m/s for water and similar liquids.

For a given heat exchange area the cheapest exchanger is one, which has a small shell
diameter and a long tube length.

Finned tubes

Stegosaurus Wuerhosaurus Loricatosaurus

104
Tube pitch
The tubes are arranged on triangular (most popular) or square orientations. The tube
pitch gives the distance between the tube centers. Unless the shell side fluid tends to
foul badly, triangular pitch is used, because about 15% more tube can be packed into a
shell of given diameter and pitch than in square pitch.

Flow Flow

Tubes in triangular pitch cannot be

cleaned by running a brush between the
rods because no space exists for cleaning
Pitch lanes. So it is suitable for clean services
or fouling services handled by chemical
treatment (not mechanical).

Square (90°) Pitch Square pitch allows mechanical cleaning

Triangular (30°) of the outside of the tubes. Also square
pitch gives a lower shell-side pressure
drop than triangular pitch.

The 90° square pitch is used in turbulent

Pitch
flow, the 45° square pitch is used in
laminar flow.
Square (45°)

Tube sheet Tube bundle with baffle

105
Fluid allocation
In principle, either stream entering a shell-and-tube heat exchanger may be put on
either tube-side or shell-side. However, there are four considerations that influence the
choice to result in the most economical exchanger.
(1) High pressure: If one of the streams is at high pressure, it should be put through
the tubes. In this case, only the tube-side fittings need to withstand the high
pressure, the shell may be made of lighter metal.
(2) Corrosion: Since most corrosion resistant alloys are more expensive than the
ordinary materials of construction, the corrosive fluid should be placed in the
tubes.
(3) Fouling: Certain streams foul so badly that the entire design is dominated by
features, which seek (a) to minimize fouling (e.g. high velocity, avoidance of
dead or eddy flow regions) (b) to facilitate cleaning (fouling fluid on tube-side,
wide pitch and rotated square layout if shell-side fluid is fouling) or (c) to extend
operational life by multiple units.
(4) Low heat transfer coefficient: If one stream has an inherent low heat transfer
coefficient (such as low pressure gases or viscous liquids), this stream should be
put on shell-side so that extended surface may be used to reduce the total cost of
the exchanger.

7.5 Energy balance

The first step in design is to set up material and energy balances over the heat
exchanger. The exchange of heat between the 2 fluids is assumed to be complete
(100%); that is the heat transfer to or from the ambient is usually neglected (negligible
in comparison with the heat transfer in the exchanger).

q = m (hout − hin ) where q = rate of heat transfer into stream; also called
heat load or heat duty, W
h = enthalpy per unit mass of stream, J/kg
m = mass flow rate of stream, kg/s

Applying it for both the hot and cold streams:

q = m h (hh,out − hh,in )
qc = − q h
q = m c (hc,out − hc,in )

m h (hh, out − hh, in ) = m c (hc, out − hc, in ) This is the overall enthalpy balance

If constant specific heats (cp) are assumed:

q = m h c p , h (Th, in − Th, out ) = m c c p , c (t c, in − t c, out )

q = m h c p , h (T1 − T2 ) = m c c p , c (t 2 − t1 )

106
 7.6 Logarithmic mean temperature difference (LMTD)

T T1
)∆T
T22
T2 tt1 Let’s consider counterflow.
2

)∆T
T11 T, t = average temperatures of hot
and cold fluid respectively, °C
t1t 2
ΔT = T − t = overall local temperature
) T1 difference, °C
∆T1
)∆T
T22
q (or A or L)

It is reasonable to expect the heat flux to be proportional to a driving force (ΔT). The
two fluid stream temperatures T and t vary with distance z within the exchanger.
Therefore the driving force for heat transfer ΔT = T − t also varies throughout the
exchanger. Consequently, it is necessary to start with a differential equation.

dq = U ΔT dA where ΔT = T − t,°C
U = overall heat transfer coefficient, W/m2K

To integrate this equation, we have to assume that:

– U is constant,
– cp,hot and cp,cold are constant,
– heat exchange with the ambient is negligible (that is there are no heat
losses),
– the flow is steady.

Since T and t vary linearly with q (assumption 2 and 4), therefore ΔT does likewise,
and d(ΔT)/dq, the slope of the graph is constant. Therefore
d (∆T ) ∆T2 − ∆T1 where qT = rate of heat transfer in entire
=
dq qT exchanger
ΔT1, ΔT2 = terminal temperature
differences or approach
d (∆T ) ∆T2 − ∆T1
=
U ∆T dA qT

∆T2 AT
d (∆T ) U (∆T2 − ∆T1 )
∫
∆T1
∆T
=
qT ∫ dA
0

∆T2 U (∆T2 − ∆T1 )

ln = AT
∆T1 qT

107
 ∆T2 − ∆T1
qT = U AT = U AT ∆Tlm
∆T2
ln
∆T1

∆T2 − ∆T1
∆Tlm =
∆T logarithmic mean temperature difference (LMTD)
ln 2
∆T1

Therefore, the heat transferred for the entire exchanger is: q = U A ΔTlm

This last equation is true not only for counterflow but parallel flow as well.

7.7 Temperature correction factor FT

In multipass shell-and-tube exchangers the flow pattern is complex (parallel,
countercurrent and crossflow are all present). A 1-2 shell-and-tube exchanger has 1
parallel pass and 1 counterflow pass. Hence, ΔTlm , which applies to either parallel or
counterflow but not to a mixture of both types, cannot be used to calculate the mean
temperature drop without a correction. It is customary to define a temperature
correction factor FT , which is so determined that when it is multiplied by the LMTD
(ΔT lm ) for counterflow, the product is the true average temperature drop:

true ΔT = FT ΔTlm

In the literature correction factor (FT) charts are available for different exchangers.

parameter: R
1.0

R
FT

0.6
0 1.0
P
Each curved line in the chart corresponds to a constant value of the dimensionless ratio
R:
T1 − T2 C
R = = c heat capacity ratio
t2 − t1 Ch

108
The abscissas are values of the dimensionless ratio P defined as:
t2 − t1
P = thermal effectiveness
T1 − t1
t2 − t1 = the actual temperature rise of the cold fluid
T1 − t1 = the maximum possible temperature rise if the warm end approach
were zero ( T1 − t2 = 0 ) based on counterflow

The temperature correction factor expresses the efficiency of an exchanger:

temperature difference in the exchanger

η= = FT
temperature difference for counterflow

In above definition the counterflow exchanger is assigned an efficiency of 100%.

Temperature correction values less than 0.75 are generally unacceptable because the
exchanger configuration chosen is inefficient. To raise the FT factor the exchanger
configuration or temperature levels, or both must be changed.

109
 FT Temperature correction factor

110
P Thermal efficiency

P=
 FT Temperature correction factor

111
P Thermal efficiency

P=
7.8 Overall heat transfer coefficient
The heat transferred for the entire exchanger is:

q = U A ΔTlm where ΔTlm = LMTD, °C

A = heat transfer area. It is customary to use the outside
surface area Ao of the inner tube (or tubes), m2
U = overall heat transfer coefficient, W/m2 K

The resistances to heat transfer are:

1 1 ln ( Ro / Ri ) ln ( Do / Di )
RTi = , RTo = , RW = =
hi Ai ho Ao 2π k L 2π k L

where RTi = film resistance for the inside tube-surface area

RTo = film resistance for the outside tube-surface area
RW = tube-wall conduction resistance
Ro = outside radius of tube
Ri = inside radius of tube
hi = film transfer coefficient based on the inside surface area A i = πDi L
ho = film transfer coefficient based on the outside surface area Ao = πDoL

It is necessary to specify the area. If A is taken as the outside tube area Ao , U becomes
a coefficient based on that area (Uo). If the inside area Ai is chosen, the coefficient is
also based on that area (Ui).
1
Uo =
1 D ln ( Do / Di ) 1 Do
+ o +
ho 2k hi Di

∆Tlm 1
q = UA∆Tlm = ⇒ UA =
Rtotal Rtotal

For thin wall pipes Ai ≅ AO = A, that is the pipe wall resistance can be neglected and

1
U=
1 1
+
ho hi

112
Note:
When not stated, the heat transfer area in conventional shell-and-tube heat exchangers
is considered as Ao outside tube surface area.

Typical values of overall heat-transfer coefficients

in shell-and-tube heat exchangers

U, W/m2K
Water to water 850-1700
Water to brine 570-1150
Water to condensing steam 570-1150
Water to petrol or kerosene 300-1000
Water to oil, gas oil or vegetable oil 100-350
Gas oil to gas oil 110-285
Steam to boiling water 1450-2250
Light organics to light organics 230-425
Heavy organics to heavy organics 55-230
Feedwater heaters 1000-8500
Steam to light fuel oil 200-400
Steam to heavy fuel oil 50-200
Gas to gas 10-40
Steam condenser 1000-6000

7.9 Fouling factors

In actual service, heat-transfer surfaces do not remain clean. Scale, dirt and other
solids deposit on one or both sides of the tubes provide additional resistances to heat
flow and reduce the overall heat transfer coefficient. In petroleum processes coke and
other substances can deposit. Corrosion products may form on the surfaces. Biological
growth (e.g. algae) can occur with cooling water and in the biological industries.
Particles held in suspension in the flow stream will deposit on the surface in low
velocity regions. Chemical reaction in the flow may take place on the heat transfer
surface producing an adhering solid product of reaction. Dissolved materials can
crystallize out under certain circumstances from the fluid on the warmer surfaces.

To avoid or reduce these fouling resistances chemical inhibitors are often added to
minimize corrosion, salt deposition, and algae growth. Water velocities above 1 m/s
are generally used to reduce fouling. Ideal velocities to minimize fouling are 1.5-2 m/s
in tubes and 1-1.5 m/s in shells. Large temperature differences should also be avoided
to reduce salt deposition on surfaces.

113
Types of fouling
‒ Chemical fouling
Chemical reaction in the fluid cause a fouling layer to be deposited onto the
tube surface (organic compounds may decompose forming deposits like
coke, tar; polymerization ...)
‒ Corrosion fouling
The heat transfer surface itself reacts to form corrosion products and a layer
builds up on the surface of the tubes.
‒ Biological fouling
Algae, bacteria or fungi growth within the fluid deposits onto the heat
exchanger surfaces forming slimes. Cooling water and some other process
streams can have this problem.
‒ Crystallization fouling
It is scale formation on heat transfer surfaces due to dissolved sulphates and
carbonates in cooling water. Salts deposit because their solubility decreases
with temperature. (CaCO3, CaSO4, Ca-oxalate, Mg(OH)2, silicates ...)
‒ Sedimentation fouling (or particulate fouling)
Suspended particles present in the fluid settle onto surfaces in low velocity
regions

114
The effect of fouling is taken care of in design by adding the fouling resistances Rdi
and Rdo to the total resistance. As a rule, they are applied without correcting for
inside diameter to outside diameter, since they are not known to any great degree of
accuracy. So the overall heat transfer coefficient becomes:

1
Uo =
1 D ln ( Do / Di ) 1 Do
+ Rdo + o + Rdi +
ho 2k hi Di

Rdi = fouling resistance or fouling factor for the inside tube surface, m2 K /W
Rdo = fouling resistance or fouling factor for the outside tube surface, m2 K /W

The fouling resistances are usually measured for various combinations of fluids and
metals. Typical fouling factors recommended for use in designing heat exchangers are
available in many references.

Approximate fouling factors for heat exchanger tubes

Material Fouling factor Rd, m2K/W

Distilled water 0.000 1
Sea water 0.000 2 - 0.000 4
Treated boiler feed water 0.000 1 - 0.000 2
Clean river or lake water 0.000 2 - 0.000 6
Steam (oil free) 0.000 1
Steam (oil-bearing) 0.000 3
Organic vapours 0.000 1
Engine oil 0.000 2
Most industrial liquids 0.000 2
Most refinery liquids 0.000 2 - 0.000 9
Gases 0.000 2 - 0.000 4
Vaporizing liquids 0.000 35
Air 0.000 4
Vegetable oils and gas oils 0.000 5
Flue gases 0.001 - 0.002
Diesel engine exhaust 0.002

Cost of fouling
Globally, fouling costs industry around € 18 billion/year ($20 billion/year). It is made
up of different cost elements:
− Additional capital cost: additional surface area is needed to compensate for
reduction in heat transfer due to fouling. An increase in exchanger size increases the
installation cost as well. A larger exchanger has increased pressure loss, therefore
larger pumps and motors are needed.
− Energy cost: increased pumping power is needed to meet increased pressure loss.
115
− Maintenance cost: Dismantling, cleaning, and reassembling, cost of chemicals and
additives to reduce fouling.
− Shutdown cost: is the value of lost production when the exchanger is shut down for
cleaning.
− Extra environmental management cost: disposing of large amount of
chemicals/additives used for fouling mitigation (hazardous or toxic cleaning
solutions or additives)

Control of fouling
Once the combination of mechanisms of fouling problem is recognized, fouling rate
may be substantially reduced. Corrosion fouling can be reduced by selecting corrosion
resistant materials for the equipment. Biofouling can be reduced by the use of Cu
alloys and/or chemical additives to the fluid stream to control the growth and
reproduction of the organisms. Particulate fouling can be reduced by ensuring
sufficient flow velocities and reducing stagnant areas.

Removal of fouling deposits

Chemical removal of fouling can be achieved in some cases by weak acid, special
solvent, and so on. Other deposits adhere weakly and can be washed off by periodic
operation at very high velocities or by flushing with a high-velocity steam or water jet
or using a sand-water slurry. These methods may be applied to both the shell side and
the tube side without pulling the bundle. Many fouling deposits however must be
removed by mechanical cleaning such as rodding, brushing, or scraping the surface.

Fluid allocation
− Fixed tube sheet: The inside of the tubes may be cleaned mechanically after
removing the channel covers or complete channels, but because the tube bundle
cannot be removed, cleaning of the outside of the tubes can only be achieved by
chemical means. Therefore, this type is limited to applications where the shell-side
fluid is virtually non-fouling; fouling fluids must be routed through the tubes.
− U-tube: The tube bundle can be withdrawn, therefore the outside of the tubes can be
cleaned mechanically, but the inside of the tubes is usually cleaned by chemical
methods.
− Corrosive fluids must always be routed through the tubes.
− The higher temperature fluid should be inside the tubes to minimize heat loss.
− If the lower flow rate fluid is allocated to the shell side, a better heat transfer
performance is achieved.
− Higher heat transfer rates are ordinarily obtained by placing a viscous fluid on the
shell-side.

7.10 Viscosity correction factor ɸv

The properties of fluid (especially viscosity) are temperature dependent. Since a
temperature field (in which the temperature varies from point to point) exist in a
flowing stream undergoing heat transfer, a problem appears in the choice of
temperature at which the properties should be evaluated.

116
One of the methods is then to evaluate the properties at the average fluid temperature
and then a correction factor (Φv) is specified.
The heat transfer coefficient h is given by

h = h′ Φ v where h’ = heat transfer coefficient calculated for Φv = 1

n
 µ 
Φ v =   = viscosity correction factor
 µw 
Liquids being heated μw < μ ⇒ Φv > 1
being cooled μw > μ ⇒ Φv < 1
For gases the inequalities are reversed.

n = 0.14 usually
μ = dynamic viscosity at mean fluid temperature, Pa⋅s
μw = dynamic viscosity at the wall temperature, Pa⋅s

Φv cannot be calculated directly because it is related to the TW surface temperature,

which in turn depends on h heat transfer coefficient. To avoid tedious trial and error
solution, the surface temperature may be calculated directly from the following
equation:
Tavg − TW
q= (
= U A Tavg − t avg )
∑ RT
⇒ TW = Tavg − Uo (Tavg − tavg )∑RT

where ∑RT = sum of appropriate thermal resistances

This TW temperature value is used to calculate Φv viscosity correction factor and the
corrected heat transfer coefficient h.

Trial-and-error calculation
Assume TW, get μw at TW, then substitute into the heat balance equation:

n
 µ  T1 − T5
ho   (T5 − T6 ) = T1
µ
 w ∑R 1− 5
●●● ● T6
or directly from T2 T5

TW = T1 – U ΔT ΣR1 – W

If it does not balance, assume a new TW. When there is a significant viscosity gradient,
Φv viscosity correction factor is no longer unity.

117
7.11 Heat transfer correlations for the shell-side
The mechanism of heat flow in forced convection outside tubes differs from that of
flow inside tubes, because of differences in fluid flow mechanism.

Inside tube: No form drag exists except perhaps for a short distance at the entrance
end, and all friction is wall friction. Therefore, there is no variation in the local heat
transfer coefficient at different points in a given circumference and a close analogy
exists between friction and heat transfer. Also, a sharp distinction exists between
laminar and turbulent flow, which calls for different treatment of heat-transfer
relations for the 2 flow regime.

Outside tube: In flow of fluids across a cylindrical shape, boundary-layer separation

occurs, and a wake develops that causes form friction. No sharp distinction is found
between laminar and turbulent flow, and a common correlation can be used for both
low and high Re numbers. Also, the local heat transfer coefficient varies from point to
point around the circumference. In practice, the variation in the local heat transfer
coefficient is often of no importance, and average values based on the entire
circumference are used.

The variables affecting the heat transfer coefficient h are the same that in case of heat
flow to fluids inside tubes: Do, cp, μ, k, and G (mass velocity of the fluid approaching
the tube).

There are many correlations and nomograms for shell-side heat transfer coefficient
calculation in literature.

118
8 CONDENSATION HEAT TRANSFER

Condensation is a convection process associated with a phase change of fluid.

Condensation occurs when a saturated vapour comes into contact with a solid surface
whose surface temperature TW is below the saturation temperature Tsat, to form a
liquid. Processes of heat transfer accompanied by phase change are more complex than
simple heat exchange between fluids.

The condensing vapour may consist of:

− a single substance (e.g. steam),
− a mixture of condensable and non-condensable substances (e.g. steam and air),
− a mixture of 2 or more condensable vapours.

The condensing temperature of a single pure substance depends only on the pressure;
therefore, the process is isothermal. Mixed vapours condensing at constant pressure
condense over a temperature range and yield a condensate of variable composition
until the entire vapour stream is condensed.

There are 2 types of condensation: – dropwise condensation,

– film-type condensation.

Dropwise condensation
The condensate begins to form at microscopic nucleation sites (e.g. tiny pits, scratches,
dust specks, ...). The drops grow and coalesce with their neighbours to form visible
fine drops. The fine drops, in turn, coalesce into small streams, which flow down the
surface under the force of gravity, sweep away condensate, and clear the surface for
more droplets. During dropwise condensation, large portion of the surface area is
covered with an extremely thin film of liquid of negligible thermal resistance;
consequently, the heat transfer coefficient at these bare areas is very high.

The average coefficient for dropwise condensation may be 5-10 times that for film-
type condensation and can be as high as 110 000 W/m2K.

Dropwise condensation occurs when the condensate does not wet the surface (e.g.
ethylene glycol, glycerine, nitrobenzene, isoheptane, steam, etc.). This type of
condensation is so unstable and is so difficult to maintain (because the surfaces
become wetted after prolonged exposure to a condensing vapour) that the method is
not common.

For normal design purposes, film-type condensation is assumed.

Film-type condensation
When the condensate wets the surface, it forms a continuous liquid film through which
heat must be transferred. The liquid film flows over the surface by the action of

119
gravity. It is this film of liquid between the surface and the vapour that forms the main
resistance to heat transfer. The thickness of the film increases rapidly in the first few
cm and then more and more slowly.

Condensation on a vertical surface (a) Dropwise (b) Film

Condensing film coefficients are much greater than those in forced convection and are
of the order of magnitude of several thousand of W/m2 K.

The film may flow in the laminar or in the turbulent regime, depending upon:
– the rate of condensation,
– fluid properties,
– length of path of the growing condensate film, and
– geometry.

The usual design for condensers employs tubes, with the condensate formed inside or
outside of the tubes. Pure vapours are usually condensed on the outside of tubes
(shell side).

The same basic design equations used for heat exchangers are valid, but the nature of
the coefficient correlations is somewhat different, depending upon whether the
condensation process takes place on vertical or horizontal tubes. Friction losses in a
condenser are normally small, so that condensation is a constant pressure process.

The basic equations for the rate of heat transfer in film-type condensation were first
developed by Nusselt (1916).

120
8.1 Vertical surfaces — Laminar flow
On a vertical flat surface of width W a vapour is condensing in a film-type manner.
The film is 2-dimensional and has a thickness of δ at any z location [δ = δ(z)].

Nusselt’s assumptions:
W (1) Pure vapor is at its Tsat saturation
temperature.
x (2) The condensate film flows in laminar flow.
y dS = W dy (3) Heat is transferred through the film solely
z by conduction ( ⇒ h z = k /δ ).
element (4) The temperature distribution through the
film is linear.
(5) The temperature of the condensing surface
dz
TW is constant.
(6) The physical properties of the condensate
y are constant and evaluated at a mean film
*δ dA = W dz
temperature Tf :
plate T + TW
T f = sat
2
(7) Negligible vapour shear exists at the
interface.

The velocity distribution in the falling film

Force balance: Buoyant force

Shear
gravity force = buoyancy + friction(shear) force force

Gravity force

ρ g (δ − y ) W dz = ρ v g (δ − y )Wdz + τ yz W dz

(δ – y) Wdz = volume of the element (control volume)

ρ = liquid density
ρv = vapour density
τyz = μ (dvz/dy) (in – z direction, laminar flow)

dv z
( ρ − ρ v ) g (δ − y ) = µ
dy

integrating

g  y 2 
vz = 
(ρ − ρv ) δ y − + C1
µ  2 
 

B.C.: at y = 0 vz = 0 ⇒ C1 = 0

121
 g y 1 y 2 
vz = (ρ − ρv ) δ 2  − velocity profile
µ δ
 2 δ 2 

m ( z )

TW

vz(y)

Mass flow rate of condensate

The differential mass flow rate through the element is ρvz dS. (dS=Wdy or dS=dydx)

The total mass flow rate at any point z is:

Wδ δ
 1 
∫∫ ∫∫ ∫
g
m = ρ v z dS = ρ v z dy dx = W ρ ( ρ − ρ v )  δ y − y 2  dy =
µ  2 
0 0 0

δ
ρ (ρ − ρv )g W  y2 y 3  Wρ ( ρ − ρ v ) g δ 3
= δ − =
µ  2 6  3µ
 0

The mass flow rate is traditionally expressed as mass flow rate per unit width of fall:

m
Γ ≡ liquid loading or condensate loading, kg/ms
W

1 /3
m δ 3 ρ ( ρ − ρv ) g  3µ Γ  for ρ >> ρv
= Γ= ⇒ δ =  2 
W 3µ ρ g

122
 m m/t ρ (W L δ )
and Γ = = = = ρ δ vavg
W W Wt

Rate of heat transfer

∂T
The heat flow at the wall is: q y = − k dA
∂y y=0

∂T ∆T TW − Tsat
Assuming a linear temperature profile: = ⇒ q y = − k W dz
∂y δ δ

In the same dz distance, the increase in mass flow rate from condensation is dm , or dΓ
for unit width:

dm δ 3 ρ ( ρ − ρ v ) g  3 δ 2 ρ (ρ − ρv ) g
= dΓ = d   = dδ
W  3µ  3 µ

Making a heat balance for dz distance:

heat added by condensation = heat removed at the wall by conduction

Tsat − TW
dm ∆H v = k Wdz
δ
the temperature drop across the condensate film:

ΔTo =Tsat – TW

ΔHv = latent heat of vaporization, kJ/kg

dm ∆T
∆H v = k dz o
W δ
substituting for dm / W

δ 2 g ( ρ − ρv ) g ∆H v ∆To
dδ = k dz
µ δ

k µ ∆To
δ 3 dδ = dz
ρ ( ρ − ρ v ) g ∆H v

integrating
δ4 k µ ∆To
= z + C2
4 ρ ( ρ − ρ v ) g ∆H v

123
 B.C.: at z = 0 δ = 0 ⇒ C2 = 0

1/ 4
 4k µ ∆To 
δ ( z ) =  z  ⇒ δ ~ z1/4
 ρ ( ρ − ρ v ) g ∆H v 

At the wall, using the local heat transfer coefficient h z, a heat balance for the element:

heat flow through the wall by conduction = heat taken away by convection

hz dA ∆To = k dA ∆To / δ

⇓
k
hz =
δ

1/ 4
 ρ ( ρ − ρ v ) g ∆H v
k 
hz = = k  
δ  4k µ ∆To z 

By integrating the local heat transfer coefficient hz over the entire surface we obtain
the average coefficient hz .
1/ 4 L 1/ 4
1 WL
1  k 3 ρ ( ρ − ρ v ) g ∆H v  1
h = ∫ ∫ hz dz dx = L   ∫  z  dz
LW  4 µ ∆To 
00   0

1/ 4 1/ 4
4  k 3 ρ ( ρ − ρ v ) g ∆H v  4  k 3 ρ ( ρ − ρ v ) g ∆H v 
h =   = hz = 0.943  
3  µ ∆To L  3 L
µ ∆To L 
   

1/ 4
 k 3 ρ ( ρ − ρ v ) g ∆H v 
⇒ h = 0.943  
 Nusselt’s equation for vertical surfaces
 µ ∆To L 

Although Nusselt’s analysis was carried out for a vertical plane surface, the result is
applicable to condensation inside or outside vertical tubes, because the condensate
film thickness δ is typically 2 or 3 order of magnitude smaller than the typical tube
diameters.

124
Nusselt’s equation gives lower h values than the
experimental values. This implies that the assumption
of laminar flow is not quite right. For laminar flow
Reδ < 1800. At values as low as Reδ = 30 – 40,
ripples begin to appear in the film surface because of
vapour shear.

Because of these turbulent effects and the effect of

vapour shear McAdams recommended 20% increase
of the h value (0.943x1.2=1.13). Calculation of h
from this equation is direct, if L is known.

McAdams further recommended a different formula

for mean film temperature:
Flow regimes during film
3 (Tsat − TW ) 3 condensation on a vertical plate
T f = Tsat − = Tsat − ∆To
4 4

When L is known:

m m where P = wetted perimeter or drainage premier

Γ ≡ = P = πD for vertical pipe
P πD

∆H v h πD h
and q = m ∆H v = h πDL ∆To ⇒ = =
L ∆To m Γ

1/ 4 1 /3 McAdams equation
 k 3 ρ ( ρ − ρ v ) g ∆H v   k 3ρ 2 g 
h = 1.13  
 = 1.18  
 laminar flow
 µ ∆To L   µ Γ  vertical tube

m m / N t
For a single tube: Γ = , for a tube bundle: Γ =
πD πD

Either part of the equation may be used, depending upon the information at hand.

The criterion for determining whether the film is laminar or turbulent is the Reynolds
number:

ρ v De ρ v 4δ 4Γ for laminar flow Re < 1800

Re = = =
µ µ µ for turbulent flow Re > 1800

125
The wetted perimeter (P), the condensation cross-sectional area (Ac), and the hydraulic
diameter or equivalent diameter (Dh) for some common geometries

8.2 Vertical surfaces — Turbulent flow

From experimental results developed by Kirkbride (1934):
1 /3
 k 3ρ 2 g   4Γ 
0.4
turbulent flow (Re > 1800)
h = 0.0077  
 µ 
 µ2   
vertical surface
 

Mixtures of vapours and non-condensing gases are usually cooled and condensed
inside vertical tubes, so that the inert gas is continually swept away from the heat
transfer surface by the incoming stream.

8.3 Horizontal surfaces

Outside tubes
The condensate film begins growing from the top of the tube with zero thickness and
increases in thickness as it flows around to the bottom and then drips off.

Nusselt’s equation for horizontal surfaces and single horizontal tube:

1/ 4 1 /3
 k 3 ρ ( ρ − ρ v ) g ∆H v   k 3ρ 2 g  outside
h = 0.725  
 = 0.954  

 µ ∆To Do   µ Γ 
horizontal surface

m
where Γ =
L

126
Horizontal tube bundle
The liquid condensate from the top tube drips onto the one below therefore, the above
equation for horizontal surface is slightly modified:
1/ 4
 k 3 ρ 2 g ∆H v 
h = 0.725  

 µ ∆To Do N 
where N = number of tubes in a vertical row (stack) [Geankoplis; Bejan]
N t = total number of tubes in the tube bank [Janna]
N ⅔ = ⅔ power is because the condensate splashes away from each
individual tube as it falls cumulatively from tube to tube instead of
dripping completely to the tube below [McCabe; Kern]

On horizontal tubes, the flow of condensate film is always laminar because of the short
path length.

In general, the coefficient of a film condensing on a horizontal tube is considerably

larger than on a vertical tube under otherwise similar conditions unless the tubes are
very short or there are very many horizontal tubes in each stack.

hhorizontal > hvertical

Inside tubes
During the condensation process, a mixture of vapour and liquid travels through the
tube. At high flow rates (turbulent flow) the liquid (in the form of droplets) and the
vapour travel together as a uniform mixture in the core of the tube, with a liquid film
condensing along the tube surface.

For low flow rates, however, vapour condenses on the tube

Vapor
surface and collects in the lower portion of the tube under the
~~~~~ action of gravity.
Condensate

ρv vv Di
For low vapour velocities Re = < 35 000 where Re is evaluated at inlet
µv conditions to the tube.

Chato equation:
1/4
 k 3 ρ 2 g ∆H v′  3
h = 0.555   where ∆H v′ = ∆H v + c p ∆To
 µ ∆To Di  8
 

127
APPENDIX

Cylindrical and spherical coordinate systems

z P (x,y,z) or P (r,Θ,z)
(r,1 ,z)
*δzz

*δ1Θ
Unit vectors and cylindrical
*δrr coordinates of point P
Position vector
z
1Θ r y
P’
Base plane
x

Differential volume element dV = r dr dΘ dz

Differential line elements dr, r dΘ, and dz
Differential surface elements (r dΘ)(dz) perpendicular to the r direction
(dz)(dr) perpendicular to the Θ direction
(dr)(r dΘ) perpendicular to the z direction

128
 z

*δrr Unit vectors and

*δkφ
P spherical coordinates
of point P
*δ1 Θ
Position vector

Θ
1
y
kφ

Base or equatorial plane

x

Differential volume element: dV = r2 sinΘ dr dΘ dΦ

Differential line elements: dr, r dΘ, and r sinΘ dΦ
Differential surface elements:
(r dΘ)( r sinΘ dΦ) perpendicular to the r direction
(r sinΘ dΦ)(dr) perpendicular to the Θ direction
(dr)(r dΘ) perpendicular to the Φ direction

129
Elemental volumes
(a) rectangular coordinates; (b) cylindrical coordinates; (c) spherical coordinates

130
Heat exchangers

Floating head heat exchanger

1. Stationary Head-Channel 21. Floating Head Cover – External
2. Stationary Head-Bonnet 22. Floating Tubesheet Skirt
3. Stationary Head Flange - Channel or Bonnet 23. Packing Box
4. Channel Cover 24. Packing
5. Stationary Head Nozzle 25. Packing Gland
6. Stationary Tubesheet 26. Lantern Ring
7. Tubes 27. Tierods and Spacers
8. Shell 28. Transverse Baffles or Support Pates
9. Shell Cover 29. Impingement Plate
10. Shell Flange - Stationary Head End 30. Longitudnal Baffle
11. Shell Flange - Rear Head End 31. Pass Partition
12. Shell Nozzle 32. Vent Connection
13. Shell Cover Flange 33. Drain Connection
14. Expansion Joint 34. Instrument Connection
15. Floating Tubesheet 35. Support Saddle
16. Floating Head Cover 36. Lifting Lug
17. Floating Head Cover Flange 37. Support Bracket
18. Floating Head Backing Device 38. Weir
19. Split Shear Ring 39. Liquid Level Connection
20. Slip-On Backing Ring

Tube sheet

131
Plate and frame heat exchanger Plate and frame heat exchanger flow pattern

gasket (ensures that the cold and hot

fluid don't mix)

Plate and frame heat exchanger plates

Herringbone or chevron plates offer a variety of
thermal lengths. They are used in a multitude of
applications and are generally more efficient with
relatively clean solutions due to greater induced
turbulence within the flow stream.

Easy to clean ⇒ food, pharmaceutical industry

U is much higher than in shell-and-tube HE
Compact design

132
 U-tube bundle

Doubble pipe heat exchanger

133
QUANTITY SI unit CONVERSION
Density (ρ) kg/m3 1g/cm3 = 1000 kg/m3
1 lb/ft 3 = 16.018 kg/m3
Energy, Work kJ 1 Btu = 1.05506 kJ
1 kcal = 4.1868 kJ
1 Btu = 252.16 cal
1 kWh = 3.6 MJ
Force (F) N 1 lbf = 4.4482 N
1 N = 10 5 dyn
Heat flux (q″) W/m2 1 Btu/ft 2 h = 3.154 W/m2
Heat rate (q) W / J/s 1 Btu/h = 0.29307 W
Heat transfer coefficient (h,U) W/m2 K 1 Btu/ft 2 h°F = 5.6783 W/m2 K
Latent heat (ΔHv) kJ/kg 1 Btu/lb = 2.326 kJ/kg
Length (L) m 1 m = 100 cm = 1000 mm
1 ft = 0.3048 m
1 in = 2.54 cm
Mass kg 1 lb = 0.4536 kg
Power W 1 hp = 0.7457 kW
Pressure (p) Pa / N/m2 1 bar = 10 5 Pa
(bar) 1 atm = 101.3 kPa
1 atm = 760 mmHg
1 atm = 14.7 psia
1 psia = 6894.76 Pa
Specific heat (cp) kJ/kgK 1 Btu/lboF = 4.1868 kJ/kgK
1 Btu/lboF = 1 kcal/kgoC
Temperature (T) K T(K) = t(oC) + 273.15
(oC) ΔT(oC) = ΔT(K) = 1.8ΔT(oF)
X(oF) = 5/(X!32)/9 (oC)
1K = 1.8oR
T(oC) = (1/1.8)(oR! 492)
Thermal conductivity (k) W/mK 1 Btu/ft hoF = 1.7307 W/mK
Viscosity Dynamic (μ) Pa≅s = kg/ms 1 cP = 10!3 Pa≅s
Kinematic (ν) m2 /s 1 cSt = 10!6 m2 /s
Volume (V) m3 1 m3 = 1000 Ρ
1 gal (US) = 3.7854 l
1 gal (GB) = 4.5435 l
1 ft 3 = 28.3168 l

134
135