Equipment Design – Fourth Year Jwan T.

Hasan

Step 6: Layout and tube size

Using a split-ring floating head exchanger for efficiency and ease of cleaning.
Neither fluid is corrosive, and the operating pressure is not high, so a plain carbon steel
can be used for the shell and tubes. The crude is dirtier than the kerosene, so put the
crude through the tubes and the kerosene in the shell.

Tube diameters: tube diameters in the range 5/8 in. (16 mm) to 2 in. (50 mm) are
used. The smaller diameters 5/8 in. to 1 in. (16 to 25 mm) are preferred for most
duties, as they will give more compact, and therefore cheaper, exchangers. Larger
tubes are easier to clean by mechanical methods and would be selected for heavily
fouling fluids. As a guide, 3/4 in. (19 mm) is a good trial diameter with which to
start design calculations.
Lengths of tubes: the preferred lengths of tubes for heat exchangers are: 6 ft. (1.83
m), 8 ft (2.44 m), 12 ft (3.66 m), 16 ft (4.88 m) 20 ft (6.10 m), 24 ft (7.32 m). The
optimum tube length to shell diameter will usually fall within the range of 5 to 10.
Tube arrangements: The tubes in an exchanger are usually arranged in an
equilateral triangular, square, or rotated square pattern; see Figure 13. The
recommended tube pitch (distance between tube centres) is 1.25 times the tube
outside diameter; and this will normally be used unless process requirements dictate
otherwise. Where a square pattern is used for ease of cleaning, the recommended
minimum clearance between the tubes is 0.25 in. (6.4 mm).

Figure 13: Tube patterns.

Use 19.05 mm (3/4 inch) outside diameter, 14.83 mm inside diameter, 5 m long
tubes (a popular size) on a triangular 23.81 mm pitch (pitch/dia. = 1.25).
15
Equipment Design – Fourth Year Jwan T. Hasan

Figure 14: Tube arrangements, showing pass-partitions in headers.

Step 7: Number of tubes

Area of heat transfer

Number of tubes, 𝑁𝑡 =
Area of one tube

𝑞
Area of heat transfer, 𝐴 = = 𝑁𝑡𝑢𝑏𝑒𝑠 × 𝑁𝑝𝑎𝑠𝑠𝑒𝑠 × 𝜋𝑑𝐿
𝑈𝐹∆𝑇𝑚

Area of one tube (neglecting thickness of tube sheets) = 𝜋𝑑𝐿 = 𝜋 × 19.05 × 10−3 × 5
= 0.2992 m2
70.89
∴ Number of tubes = = 237, say 240
0.2992

16
Equipment Design – Fourth Year Jwan T. Hasan

240
So, for 2 passes, tubes per pass = = 120
2

Check the tube-side velocity at this stage to see if it looks reasonable.

𝜋 2 𝜋
Tube cross sectional area = 𝑑 = (14.83 × 10−3 )2 = 0.0001727 m2
4 4
𝜋 2
So, area per pass = 𝑁𝑡 × 𝑑 = 120 × 0.0001727 = 0.02073 m2
4

ṁ 70000 1
Volmetric flow, 𝑄 = = × = 0.0237 m3 /s
𝜌 3600 820
0.0237
Tube side velocity, 𝑢𝑡 = = 1.14 m/s
0.02073

The velocity is satisfactory, between 1 to 2 m/s, but may be a little low. This will
show up when the pressure drop is calculated.

Step 8: Bundle and shell diameter

An estimate of the bundle diameter 𝐷𝑏 , can be obtained from equation which is an

empirical equation based on standard tube layouts:

𝑁𝑡 1⁄n1
𝐷𝑏 = 𝑑𝑜 ( )
𝐾1

Where, 𝑁𝑡 = number of tubes,

𝐷𝑏 = bundle diameter, mm, 𝑑𝑜 = tube outside diameter, mm.
The constants for use in this equation, for triangular and square patterns, are given
in Table 2.
Table 2: Constants for the equation of bundle diameter estimation.

17
Equipment Design – Fourth Year Jwan T. Hasan

From Table 2, for 2 tube passes, 𝐾1 = 0.249, n1 = 2.207

240 1/2.207
So, Bundle diameter: 𝐷𝑏 = 20 × ( ) = 428 mm (0.43 m)
0.249

The shell diameter must be selected to give as close a fit to the tube bundle as is
practical; to reduce bypassing round the outside of the bundle. The clearance required
between the outermost tubes in the bundle and the shell inside diameter will depend on
the type of exchanger and the manufacturing tolerances; typical values are given in
Figure 15.

Figure 15: Shell-bundle clearance.

For a split-ring floating head exchanger the typical shell clearance from Figure 15
is 56 mm, so the shell inside diameter,
Shell diameter, 𝐷𝑠 = 428 + 56 = 484 mm

18
Equipment Design – Fourth Year Jwan T. Hasan

Step 9: Tube-side heat transfer coefficient

Turbulent flow: Heat-transfer data for turbulent flow inside conduits of uniform cross-
section are usually correlated by an equation of the form:

𝜇 𝑐
𝑁𝑢 = 𝐶 𝑅𝑒 𝑎 𝑃𝑟 𝑏 ( )
𝜇𝑤
ℎ𝑖 𝑑 𝑒
Where, 𝑁𝑢 = Nusselt number = ,
𝑘𝑓
𝐺𝑡 𝑑𝑒
𝑅𝑒 = Reynolds number = ,
𝜇
𝐶𝑝 𝜇
𝑃𝑟 = Prandtl number = ,
𝑘𝑓

and: ℎ𝑖 = inside coefficient, W/m2 . ℃

𝑑𝑒 = equivalent (or hydraulic mean) diameter, m

4 × cross sectional area for flow

𝑑𝑒 = = 𝑑𝑖 for tubes,
wetted perimeter
𝑢𝑡 = fluid velocity, m/s,
𝑘𝑓 = fluid thermal conductivity, W/m. ℃,
𝐺𝑡 = mass velocity, mass flow per unit area, kg/m2 . s,
𝜇 = fluid viscosity at the bulk fluid temperature, N. s/m2 ,
𝜇𝑤 = fluid viscosity at the wall,
𝐶𝑝 = fluid specific heat, heat capacity, J/kg. ℃.

The index for the Reynolds number is generally taken as 0.8. That for the Prandtl
number can range from 0.3 for cooling to 0.4 for heating. The index for the viscosity
factor is normally taken as 0.14 for flow in tubes. A general equation that can be used
for exchanger design is:
𝜇 0.14
𝑁𝑢 = 𝐶 𝑅𝑒 0.8 𝑃𝑟 0.33 ( )
𝜇𝑤
Where, 𝐶 = 0.021 for gases,
= 0.023 for non-viscous liquids,
= 0.027 for viscous liquids.

19
Equipment Design – Fourth Year Jwan T. Hasan

Laminar flow: the following equation can be used to estimate the film heat-transfer
coefficient:
0.14
0.33
𝑑𝑒 0.33 𝜇
𝑁𝑢 = 1.86 (𝑅𝑒Pr) ( ) ( )
𝐿 𝜇𝑤

Where, 𝐿 = is the length of the tube in metres.

heat transfer factor, 𝒋𝒉 : It is often convenient to correlate heat-transfer data in terms

of a heat transfer factor, which is similar to the friction factor used for pressure drop.

ℎ𝑖 𝑑 𝑖 𝜇 0.14
𝑁𝑢 = = 𝑗ℎ 𝑅𝑒𝑃𝑟 0.33 ( )
𝑘𝑓 𝜇𝑤

The use of the 𝑗ℎ factor enables data for laminar and turbulent flow to be represented
on the same graph; Figure 16, The 𝑗ℎ values obtained from Figure 16 can be used to
estimate the heat-transfer coefficients for heat-exchanger tubes and commercial pipes

So,
𝐺𝑡 𝑑𝑒 820 × 1.14 × 14.83 × 10−3
𝑅𝑒 = = −3
= 4332, (4.3 × 103 )
𝜇 3.2 × 10

𝐶𝑝 𝜇 2.05 × 103 × 3.2 × 10−3

𝑃𝑟 = = = 48.96
𝑘𝑓 0.134

𝐿 5000
= = 337
𝑑𝑖 14.83

From figure 16 , 𝑗ℎ = 3.2 × 10−3

ℎ𝑖 𝑑 𝑖 0.33
𝜇 0.14
𝑁𝑢 = = 𝑗ℎ 𝑅𝑒𝑃𝑟 ( )
𝑘𝑓 𝜇𝑤

𝑁𝑢 = 3.2 × 10−3 (4332) × (48.96)0.33 = 50.06

0.134
ℎ𝑖 = 50.06 × ( ) = 452 W/m2 . ℃
14.83 × 10−3

20
Equipment Design – Fourth Year Jwan T. Hasan

This is clearly too low if 𝑈𝑜 is to be 300 W/m2 . ℃. The tube-side velocity did look
low, so increase the number of tube passes to 4. This will halve the cross-sectional area
in each pass and double the velocity.

New 𝑢𝑡 = 2 × 1.14 = 2.3 m/s

and 𝑅𝑒 = 2 × 1.14 = 8664 (8.7 × 103 )

𝑗ℎ = 3.8 × 10−3

0.134
ℎ𝑖 = ( −3
) × 3.8 × 10−3 × 8664 × (48.96)0.33 = 1074 W/m2 . ℃
14.83 × 10

Figure 16: Tube-side heat transfer factor.

21
Equipment Design – Fourth Year Jwan T. Hasan

Step 10: Shell-side heat transfer coefficient

Kern's method will be used. With 4 tube passes the shell diameter will be larger than
that calculated for 2 passes.

For 4 passes , 𝐾1 = 0.175, n1 = 2.285.

240 1/2.285
𝐷𝑏 = 19.05 × ( ) = 450 mm, (0.45 m)
0.175

The bundle to shell clearance is still around 56 mm, giving:

𝐷𝑠 = 506 mm (about 20 inches)

As a first trial take the baffle spacing = 𝐷𝑠 ⁄5, say 100 mm. This spacing should
give good heat transfer without too high a pressure drop.

Procedure
1. Calculate the area for cross-flow 𝐴𝑠 for the hypothetical row of tubes at the shell
equator, given by:

(𝑝𝑡 − 𝑑𝑜 )𝐷𝑠 𝑙𝐵
𝐴𝑠 =
𝑝𝑡
where, 𝑝𝑡 = tube pitch,
𝑑𝑜 = tube outside diameter,
𝐷𝑠 = shell inside diameter, m,
𝑙𝐵 = baffle spacing, m.
(𝑝𝑡 −𝑑𝑜 )
The term is the ratio of the clearance between tubes and the total distance
𝑝𝑡
between tube centres.

(23.81 − 19.05)
Cross flow area, 𝐴𝑠 = 506 × 100 = 10116 mm2 = 0.01012 m2
23.81

2. Calculate the shell-side mass velocity and the linear velocity.

20000 1
Volumetric flow rate on shell side = × = 0.0076 m3 /s
3600 730

22
Equipment Design – Fourth Year Jwan T. Hasan

0.076
Shell side velocity, 𝑢𝑠 = = 0.75 m/s
0.01012

3. Calculate the shell-side equivalent diameter (hydraulic diameter). For a square pitch
arrangement:
𝑝2 − 𝜋𝑑𝑜2
4( 𝑡 ) 1.27
4
𝑑𝑒 = = (𝑝𝑡2 − 0.785 𝑑𝑜2 )
𝜋𝑑𝑜 𝑑𝑜
For an equilateral triangular pitch arrangement:

𝑝𝑡 1 𝜋𝑑𝑜2
× 0.87 𝑝𝑡 ×
4(2 2 4 )
2
1.10 2
𝑑𝑒 = = (𝑝𝑡 − 0.917𝑑𝑜2 )
𝜋𝑑𝑜 𝑑𝑜
2
where 𝑑𝑒 = equivalent diameter, m.

1.10
Equivalent diameter, 𝑑𝑒 = (23.812 − 0.917 × 19.052 ) = 13.52 mm
19.05

4. Calculate the shell-side Reynolds number,

730 × 0.75 × 13.52 × 10−3

𝑅𝑒 = −3
= 17214, (1.72 × 104 )
0.43 × 10

2.47 × 103 × 0.43 × 10−3

𝑃𝑟 = = 8.05
0.132

5. For the calculated Reynolds number, read the value of 𝑗ℎ from Figure 17, for the
selected baffle cut and tube arrangement, and calculate the shell-side heat transfer
coefficient ℎ𝑠 from:

ℎ𝑠 𝑑 𝑒 𝜇 0.14
𝑁𝑢 = = 𝑗ℎ 𝑅𝑒𝑃𝑟 0.33 ( )
𝑘𝑓 𝜇𝑤

Use segmental baffles with a 25% cut. This should give a reasonable heat transfer
coefficient without too large a pressure drop.

From Figure 17, 𝑗ℎ = 4.52 × 10−3

23
Equipment Design – Fourth Year Jwan T. Hasan

Neglecting the viscosity correction:

0.132
ℎ𝑠 = ( × 103 ) × 4.52 × 10−3 × 17214 × 8.050.33 = 1505 W/m2 . ℃
13.52

Figure 17: Shell-side heat transfer factors, segmental baffles.

Step 11: Overall coefficient

𝑑𝑜
1 1 1 𝑑𝑜 ln ( ) 𝑑𝑜 1 𝑑𝑜 1
𝑑𝑖
= + + + × + ×
𝑈𝑜 ℎ𝑜 ℎ𝑜𝑑 2𝑘𝑤 𝑑𝑖 ℎ𝑖𝑑 𝑑𝑖 ℎ𝑖

Where,

𝑈𝑜 = the overall coefficient based on the outside area of the tube, W/m2. ℃,
ℎ𝑜 = outside fluid film coefficient, W/m2. ℃,
ℎ𝑖 = inside fluid film coefficient, W/m2. ℃,
ℎ𝑜𝑑 = outside dirt coefficient (fouling factor), W/m2. ℃,
ℎ𝑖𝑑 = inside dirt coefficient, W/m2. ℃,
𝑘𝑤 = thermal conductivity of the tube wall material, W/m. ℃,

24
Equipment Design – Fourth Year Jwan T. Hasan

𝑑𝑖 = tube inside diameter, m,

𝑑𝑜 = tube outside diameter, m.

−3 19.05
1 1 19.05 19.05 × 10 ln (14.83) 1
=( + 0.00035) + + + 0.0002
𝑈𝑜 1074 14.83 2 × 55 1505
= 386 W/m2 . ℃

This is above the initial estimate of 300 W/m2 . ℃. The number of tubes could
possibly be reduced, but first check the pressure drops.

Step 12: Pressure drop

Tube-side

There are two major sources of pressure loss on the tube-side of a shell and tube
exchanger: the friction loss in the tubes and the losses due to the sudden contraction and
expansion and flow reversals that the fluid experiences in flow through the tube
arrangement

𝐿 𝜇 −𝑚 𝜌𝑢𝑡2
∆𝑃𝑡 = 𝑁𝑝 (8 𝑗𝑓 ( ) ( ) + 2.5)
𝑑𝑖 𝜇𝑤 2

𝑚 = 0.25 for laminar flow, Re < 2100,

= 0.14 for turbulent flow, Re > 2100.

Values of 𝑗𝑓 for heat exchanger tubes can be obtained from Figure 18.

240 tubes, 4 passes, tube i.d. 14.83 mm, 𝑢𝑡 2.3 m/s, 𝑅𝑒 = 8.7× 103 .
From Figure 18, 𝑗𝑓 = 5 × 10−3

−3
5000 (820 × 2.32 ) (820 × 2.32 )
∆𝑃𝑡 = 4 (8 × 5 × 10 ( ) + 2.5) = 4(13.5 + 2.5)
14.83 2 2
= 138810 N/m2 , 1.4 bar

This exceeds the specification. Return to step 6 and modify the design.

25
Equipment Design – Fourth Year Jwan T. Hasan

Figure 18: Tube-side friction factors.

Shell-side

For the calculated shell-side Reynolds number, read the friction factor from Figure
19 and calculate the shell-side pressure drop from:

−0.14
𝐷𝑠 𝐿 𝜌𝑢𝑠2 𝜇
∆𝑃𝑠 = 8 𝑗𝑓 ( ) ( ) ( )
𝑑𝑒 𝑙𝐵 2 𝜇𝑤

Where, 𝐿 = tube length,

𝑙𝐵 = baffle spacing.
𝐿
The term ( ) is the number of times the flow crosses the tube bundle = (𝑁𝑏 + 1),
𝑙𝐵

where 𝑁𝑏 , is the number of baffles.

26
Equipment Design – Fourth Year Jwan T. Hasan

Figure 19: Tube-side friction factors, segmental baffles.

Modified design

The tube velocity needs to be reduced. This will reduce the heat transfer coefficient,
so the number of tubes must be increased to compensate. There will be a pressure drop
across the inlet and outlet nozzles. Allow 0.1 bar for this, a typical figure (about 15%
of the total); which leaves 0.7 bar across the tubes. Pressure drop is roughly proportional
to the square of the velocity and ut is proportional to the number of tubes per pass. So
the pressure drop calculated for 240 tubes can be used to estimate the number of tubes
required.
Tubes needed = = 240/(0.6⁄1.4)0.5 = 365
Say, 360 with 4 passes.
Retain 4 passes as the heat transfer coefficient will be too low with 2 passes.
Second trial design: 360 tubes 19.05 mm o.d., 14.83 mm i.d., 5 m long, triangular pitch
23.81 mm.

27
Equipment Design – Fourth Year Jwan T. Hasan

360 1/2.285
Bundle diameter: 𝐷𝑏 = 20 × ( ) = 537 mm (0.54 m)
0.175

From Figure 15 clearance with this bundle diameter = 59 mm

shell diameter, 𝐷𝑠 = 537 + 59 = 596 mm.

360 𝜋
cross − sectional area per pass = × (14.83 × 10−3 )2 = 0.01555 m2
4 4
0.0237
tube velocity, 𝑢𝑡 = = 1.524 m/s
0.01555

820 × 1.524 × 14.83 × 10−3

𝑅𝑒 = = 5792
3.2 × 10−3

𝐿/𝑑 is the same as the first trial, 337

𝑗ℎ = 3.6 × 10−3

0.134
ℎ𝑠 = ( × 10−3 ) × 3.6 × 10−3 × 5792 × 48.960.33 = 680 W/m2 . ℃
14.83

This looks satisfactory, but check the pressure drop before doing the shell-side
calculation.

𝑗𝑓 = 5.5 × 10−3

−3
5000 (820 × 1.5242 )
∆𝑃𝑡 = 4 (8 × 5.5 × 10 ( ) + 2.5) =
14.83 2
= 66029 N/m2 , 0.66 bar

Well within specification.

Keep the same baffle cut and spacing

(23.81 − 19.05)
Cross flow area, 𝐴𝑠 = 596 × 100 = 11915 mm2 = 0.01192 m2
23.81
0.0076
𝑢𝑠 = = 0.638 m/s
0.01193
𝑑𝑒 = 13.52 mm, as before
730 × 0.638 × 13.52 × 10−3
𝑅𝑒 = = 14644, (1.5 × 104 )
0.43 × 10−3

28
Equipment Design – Fourth Year Jwan T. Hasan

𝑃𝑟 = 8.05

𝑗ℎ = 4.8 × 10−3 , 𝑗𝑓 = 4.6 × 10−3

0.132 −3 0.33
W
ℎ𝑠 = ( ) × 4.8 × 10 × 14644 × 8.05 = 1366 . ℃, looks OK
13.52 × 10−3 m2

596 5000 730 × 0.6382 𝑁

∆𝑃𝑠 = 8 × 4.6 ( )( ) = 120510 2 , 1.2 bar
13.52 100 2 m

Too high; the specification only allowed 0.8 overall, including the loss over the nozzles.
Check the overall coefficient to see if there is room to modify the shell-side design.
−3 19.05
1 1 19.05 19.05 × 10 ln (14.83) 1
=( + 0.00035) + + + 0.0002
𝑈𝑜 683 14.83 2 × 55 1366
= 302 W/m2 . ℃
𝑄
𝑈𝑜 required = , 𝐴𝑜 = 360 × 0.2992 = 107.7 m2
𝐴𝑜 𝐹∆𝑇𝑚

1509.4 × 103
𝑆𝑜, 𝑈𝑜 required = = 197 W/m2 . ℃
107.7 × 71

The estimated overall coefficient is well above that required for design, 302 compared
to 192 W/m2 . ℃, which gives scope for reducing the shell-side pressure drop. Allow a
drop of 0.1 bar for the shell inlet and outlet nozzles, leaving 0.7 bar for the shell-side
flow. So, to keep within the specification, the shell-side velocity will have to be reduced
by around √(1/2) = 0.707. To achieve this the baffle spacing will need to be increased
to 100/0.707 = 141, say 140 mm.

(23.81 − 19.05)
𝐴𝑠 = 596 × 140 = 6681 mm2 = 0.167 m2
23.81
0.0076
𝑢𝑠 = = 0.455 m/s
0.0167

Giving: Re = 10,443, ℎ𝑠 =1177 W/m2 . ℃, ∆𝑃𝑠 = 0.47 bar, and 𝑈𝑜 = 288 W/m2 . ℃.
The pressure drop is now well within the specification.

29