Designing Shell-and-Tube Heat Exchangers Using Softwares

Lecture 1: Kern’s Method

By:
Majid Hayati

University of Kashan, Kashan, I.R. IRAN

2014
 Schedule – Kern’s Method

Kern’s method

Introduction to Kern’s method

Algorithm of design procedure for shell-and-tube heat exchangers

Design procedure steps along with an example

3
 Objectives

This lecture on designing shell-and-tube HEs serves as an

introduction lecture to the subject, and covers:

Introduction to “Kern’s method” definition along with its

advantages and disadvantages

Developing an algorithm for the design of shell-and-tube

exchangers

Finally, following up the procedure set out in the algorithm in an

example

4
 Introduction to Kern’s method

Kern’s was based on experimental work on commercial

exchanger
Advantages:
Giving reasonably satisfactory prediction of the heat-transfer
coefficient for standard design
Simple to apply
Accurate enough for preliminary design calculations
Accurate enough for designs when uncertainty in other design
parameter is such that the use of more elaborate method is not
justified
Disadvantage:
The prediction of pressure drop is less satisfactory, as pressure
drop is more affected by leakage and bypassing than heat transfer
The method does not take account of the bypass and leakage
streams
5
 Design procedure for shell-and-tube heat exchangers
(Kern’s method)
Start from step 3 End
No
Start Yes
Compare to estimated overall
Accept all design parameters
Step 1 heat transfer coefficient

Collect physical properties and Step 8

HE specifications Determine overall heat transfer Estimate tube- and shell-side heat
coefficient transfer coefficient-go to step 3
Step 2
Define duty
Question: Are pressure drops
Make energy balance if needed Yes No
within specification?

Step 7
Calculate unspecified flow rates
Estimate tube- and shell-side
Calculate ΔTLMTD and ΔTM
pressure drop
Step 3 Step 6
Assume value of overall Estimate tube- and shell-side
coefficient Uo,ass heat transfer coefficient

Step 4 Step 5
Calculate tube number
Determine fouling factors
Calculate shell diameter
Fig. 1: Algorithm of design procedure
6
 Kern’s Method Design Example

Design an exchanger to sub-cool condensate from a

methanol condenser from 95 °C to 40 °C

Flow-rate of methanol 100,000 kg/h

Brackish water (seawater) will be used as the coolant, with a

temperature rise from 25° to 40 °C

7
 Solution: Step 1

Collect physical properties and HE specifications:

Physical properties
Table 1
Physical properties at Methanol Water
fluid mean temperature

Cp (Kj/Kg °C) 2.84 4.2

μ (mNs/m2) 0.34 0.8
kf (W/m °C) 0.19 0.59
ρ (Kg/m3) 750 995

HE specifications:
Coolant (brackish water) is corrosive, so assign to tube-side.
Use one shell pass and two tube passes.
At shell side, fluid (methanol) is relatively clean. So, use 1.25 triangular pitch
(pitch: distance between tube centers).
8
 Tube Arrangements

The tubes in an exchanger are usually arranged in an

equilateral triangular, square, or rotated square pattern (Fig.
2)

Fig. 2: Tube patterns

9
 Tube Pattern Applications

The triangular and rotated square patterns give higher heat-

transfer rates, but at the expense of a higher pressure drop
than the square pattern.
A square, or rotated square arrangement, is used for
heavily fouling fluids, where it is necessary to mechanically
clean the outside of the tubes.

The recommended tube pitch is 1.25 times the tube outside

diameter; and this will normally be used unless process
requirements dictate otherwise.

10
 Step 2

Define duty, Make energy balance if needed

To start step 2, the duty (heat transfer rate) of methanol (the hot stream or
water, the cold stream) needed to be calculated.

Fig. 3: Streams definitions.

• 100000
Heat load = Q = m Cph (T1 - T2 ) = × 2.84 (95 - 40) = 4340 kW
h 3600
11
 Step 2 (Cont’d)

The cold and the hot stream heat loads are equal. So, cooling water flow rate
is calculated as follow:
. Q 4340
Cooling water flow = mc = = _
= 68.9 kg/s
CP c (t 2 _ t1) 4.2 (40 25)

The well-known “logarithmic mean” temperature difference (LMTD or lm) is

calculated by:
(T1  t 2 )  (T2  t1 ) (95  40)  (40  25)
ΔTLMTD    31 C
T t (95  40)
ln 1 2 ln
T2  t1 (40  25)

12
 Mean Temperature Difference

The usual practice in the design of shell and tube

exchangers is to estimate the “true temperature difference”
from the logarithmic mean temperature by applying a
correction factor to allow for the departure from true
counter-current flow:
ΔTm  Ft ΔTLMTD
Where:

ΔTm = true temperature difference,

Ft = the temperature correction factor.

13
 Temperature Correction Factor

The correction factor (Ft) is a function of the shell and tube

fluid temperatures, and the number of tube and shell
passes.
It is normally correlated as a function of two dimensionless
temperature ratios:
T1  T2
R
t 2  t1

t 2  t1
S
T1  t1

14
 Step 2

For a 1 shell : 2 tube pass exchanger, the correction factor

is plotted in Fig. 4.

Fig. 4: Temperature correction factor: one shell pass; two or more even tube passes (available in TEMA)
15
 Step 2 (Cont’d)

T1  T2 95  40
R   3.67
t 2  t1 40  25

(t 2  t1 ) 40  25
S   0.21
(T1  t1 ) 95  25

From Fig. 4, the correction factor (Ft) is 0.85.

ΔTm  Ft ΔTLMTD  0.85  31  26  C

16
 Step 3

Assume value of overall coefficient Uo,ass

Typical values of the overall heat-transfer coefficient for various
types of heat exchanger are given in Table 1.

Fig. 5 can be used to estimate the overall coefficient for tubular

exchangers (shell and tube).

The film coefficients given in Fig. 5 include an allowance for fouling.

The values given in Table 1 and Fig. 5 can be used for the preliminary
sizing of equipment for process evaluation, and as trial values for
starting a detailed thermal design.

From Table 2 or Fig. 5: U=600 W/m2°C

17
 Step 3 (Cont’d)
Table 2: Typical overall coefficients

18
 Step 3 (Cont’d)

Fig. 5: Overall coefficients (join process side duty to service side and read U from centre scale)
19
 Step 4

Calculate tube number, Calculate shell diameter

Provisional area:
Q 4340 × 10 3
A= = = 278 m2
U ΔTM 600 × 26
So, the total outside surface area of tubes is 278 m2

Choose 20 mm o.d. (outside diameter), 16 mm i.d. (inside diameter),

3
4.88-m-long tubes ( in.  16 ft ), cupro-nickel.
4
Allowing for tube-sheet thickness, take tube length: L= 4.83 m

Surface area of one tube: A = πDL = 4.83 x 20 x 10-3π = 0.303 m2

Total outside surface area of tubes (Provisional area) 278

Numbers of tubes    918
Outside surface area of one tube 0.303

20
 Step 4 (Cont’d)

An estimate of the bundle diameter Db can be obtained from

equation below which is an empirical equation based on standard
tube layouts. The constants for use in this equation, for triangular
and square patterns, are given in Table 3.

N t 1n1
Db  do ( )
K1
where Db = bundle diameter in mm, do = tube outside diameter in
mm., Nt = number of tubes.
As the shell-side fluid is relatively clean use 1.25 triangular pitch.
So, for this example:
918 1
Bundle diameter D b  20 ( ) 2.207  826 mm
0.249
21
Step 4 (Cont’d)

Table 3: Constants K1 and n1

22
 Step 4 (Cont’d)

Use a split-ring floating head type for Fig. 6.

From Fig. 6, bundle diametrical clearance is 68 mm.

Shell diameter (Ds):

Ds= Bundle diameter + Clearance = 826 + 68 = 894 mm.

Note 1: nearest standard pipe size are 863.6 or 914.4 mm.

Note 2: Shell size could be read from standard tube count tables
[Kern (1950), Ludwig (2001), Perry et al. (1997), and Saunders (1988)].

23
Step 4 (Cont’d)

Fig. 6: Shell-bundle clearance

24
 Step 6

Estimate tube- and shell-side heat transfer coefficient

Tube-side heat transfer coefficient:

40 + 25
Mean water temperature (Tavg ) = = 33 C ⇒ ρ = 995 kg m3
2
π π
Tube cross - sectional area (a) = D2 = × 16 2 = 201 mm2
4 4

Since we have two tubes pass, we divide the total numbers of tubes
by two to find the numbers of tubes per pass, that is:
918
Tubes per pass = = 459
2
Total flow area is equal to numbers of tubes per pass multiply by
tube cross sectional area:
Total flow area = 459 × (201 × 106 ) = 0.092 m2
25
 Step 6 (Cont’d)

Fig. 7: Equivalent diameter, cross-sectional areas and wetted perimeters.

26
 Step 6 (Cont’d)
Cooling water flow 68.9
Water mass velocity = = = 749 kg s m2
Total flow area 0.092
Water mass velocity (Gt ) 749
Water linear velocity (ut ) = = = 0.75 m s
Water density (ρ) 995

Coefficients for water: a more accurate estimate can be made by

using equations developed specifically for water.
The physical properties are conveniently incorporated into the
correlation. The equation below has been adapted from data given by
Eagle and Ferguson (1930):
4200 (1.35 + 0.02t) u 0.8
hi =
di0.2
where hi = inside coefficient, for water, W/m2 °C,
t = water temperature, °C,
ut = water linear velocity, m/s,
di = tube inside diameter, mm.
27
 Step 6 (Cont’d)
4200 (1.35 + 0.02t) u 0.8 4200 (1.35 + 0.02 × 33) 0.75 0.8
hi = 0.2 = = 3852 W/m2  C
di 16 0.2
The equation can also be calculated using equation below; this is
done to illustrate use of this method.
hi di 0.33 μ 0.14
= jh Re Pr ( )
kf μw
where hi = inside coefficient, for water, W/m2 °C,
di = tube inside diameter, mm
kf = fluid thermal conductivity, W/m2 °C
jh = heat transfer factor, dimensionless
Re = Reynolds number, dimensionless
Pr = Prandtl number, dimensionless
μ = viscosity of water, N s/m2
μw = viscosity of water at wall temperature, N s/m2
28
 Step 6 (Cont’d)
Viscosity of water (μ) from Table 1 = 0.8 mNs m2
Fluid thermal conductivity from Table 1 = 0.59 W m C
ρudi 995 × 0.75 × 16 × 10 3
Re = = _ = 14925
μ 8 × 10 3
_
Cpμ 4.2 × 103 × 0.8 × 10 3
Pr = = = 5.7
kf 0.59
μ
Neglect (μ )
w

L 4.83 × 103 _
= = 302 ⇒ From Fig. 8, jh = 3.9 × 10 3
di 16
kf μ 0.59 _
2
hi = jh Re Pr 0.33 ( )0.14 = _ × 3.9 × 10 3
× 14925 × 5.7 0.33
× 1 0.14
= 3812 W m C
di μw 16 × 10 3

Check reasonably the previously calculated value 3812 W/m2°C with

value calculated, 3852 W/m2°C.
29
 Step 6 (Cont’d)

Fig. 8: Tube-side heat-transfer factor

30
 Step 6 (Cont’d)
Shell-side heat transfer coefficient:
Baffle spacing: The baffle spacings used range from 0.2 to 1.0 shell
diameters.
A close baffle spacing will give higher heat transfer coefficients but at the
expense of higher pressure drop.
Area for cross-flow: calculate the area for cross-flow As for the hypothetical
row at the shell equator, given by:
(pt _ do )Dslb
As =
pt
Where pt = tube pitch (distance between the centers of two tubes, Fig. 7).
do = tube outside diameter, m,
Ds = shell inside diameter, m,
lb = baffle spacing, m.
_
(p d )
t o
Note: the term p is the ratio of the clearance between tubes and
t

the total distance between tube centers.

31
 Step 6 (Cont’d)

Baffle spacing:
Choose baffle spacing = 0.2 Ds=0.2 894 = 178 mm

Tube pitch:
Pt = 1.25 do= 1.25 20 = 25 mm

Cross-flow area:

(p t _ do ) (25 _ 20) _
As = Dslb = × 894 × 178 × 10 6 = 0.032 m2
pt 25

32
 Step 6 (Cont’d)
Shell-side mass velocity Gs and the linear velocity ut:
Ws
Gs =
As
G
us = s
ρ
Where Ws = fluid flow-rate on the shell-side, kg/s,
ρ = shell-side fluid density, kg/m3.

Shell equivalent diameter (hydraulic diameter): calculate the shell-

side equivalent diameter, see Fig. 7. For an equilateral triangular
pitch arrangement: 2
pt 1 d
4( × 0.87p t _ π o ) 1.10
de = 2 2 4 = (p2t _
0.917 d2o )
πdo do
2
Where de = equivalent diameter, m.

33
 Step 6 (Cont’d)

Shell-side mass velocity Gs:

Ws 100000 1 kg
Mass velocity, Gs = = × = 868
As 3600 0.032 s m2

Shell equivalent diameter (hydraulic diameter):

1.10 2 _ 1.1
de = (p t 0.917 d2o ) = (25 2 _
0.917 × 202 ) = 14.4 mm
do 20

34
 Step 6 (Cont’d)
95 + 40
Mean shell side temperatur e = = 68  C
2
Methanol density (ρ) from Table 1 = 750 kg m3 )
Viscosity of methanol (μ from Table 1 = 0.34 mNs m2 )
Heat capacity from Table 1 = 2.84 kJ kgC
Thermal conductivi ty Table 1 = 0.19 W mC
ρu sde G sde 868 × 14.4 × 10 - 3
Re = = = _ = 36762
μ μ 0.34 × 10 3
_
Cpμ 2.84 × 10 × 0.34 × 10 3
3
Pr = = = 5.1
kf 0.19

Choose 25 per cent baffle cut, from Fig. 9

_
3
jh = 3.3 × 10

35
 Step 6 (Cont’d)

Fig. 9: Shell-side heat-transfer factors, segmental baffles

36
 Step 6 (Cont’d)

For the calculated Reynolds number, the read value of jh from Fig. 9
for 25 per cent baffle cut and the tube arrangement, we can now
calculate the shell-side heat transfer coefficient hs from:
h sde μ 0.14 0.19
Nu = = jh Re Pr 1 3 ( ) (without viscosity correction term) → hs = - 3 × 3.3 × 10 - 3 × 36762 × 5.1 1 3
kf μw 1.44 × 10
= 2740

The tube wall temperature can be estimated using the following

method:
Mean temperature difference across all resistance: 68 -33 =35 °C
U 600
across methanol film = × ΔT = × 35 = 8 C
ho 2740

Mean wall temperature = 68 – 8 = 60 °C

μ = 0.37 mNs/m2
μ 0.14
) = 0.99 (
μw
Which shows that the correction for low-viscosity fluid is not significant.
37
 Step 7 (Cont’d)

Pressure drop
Tube side: From Fig. 10, for Re = 14925
jf = 4.3 10-3
Neglecting the viscosity correction term:
L μ -m ρu2t
ΔPt = Np [8jf ( )( ) + 2.5]
di μw
3
- 3 4.83 × 10 995 × 0.75 2
= 2 (8 × 4.3 × 10 ( ) + 2.5)
16 2
2
= 7211N m = 7.2 kPa (1.1 psi)
low, could consider increasing the number of tube passes.
Shell side
G 868
Linear velocity = s = = 1.16 m/s
ρ 750
From Fig. 11, for Re = 36762
jf = 4 10-2

Neglect viscosity correction

38
 Step 7 (Cont’d)

Fig. 10: Tube-side friction factors

39
 Step 7 (Cont’d)

Fig. 11: Shell-side friction factors, segmental baffles

40
 Step 7 (Cont’d)

Ds L ρu2t - 2 894 4.83 × 103 750 × 1.162

ΔPs = 8jf ( )( ) = 8 × 4 × 10 ( )( )
de L s 2 14.4 178 2
= 272019 N m2
= 272 kPa (39 psi) too high,

could be reduced by increasing the baffle pitch. Doubling the pitch halves
the shell side velocity, which reduces the pressure drop by a factor of
approximately (1/2)2
272
ΔPs = = 68 kPa (10psi), acceptable
4
This will reduce the shell-side heat-transfer coefficient by a factor of
(1/2)0.8(ho∝ Re0.8 ∝ us0.8)
ho = 2740 (1/2)0.8 = 1573 W/m2°C

This gives an overall coefficient of 615 W/m2°C – still above assumed value of
600 W/m2°C

41
 Step 8 (Cont’d)

Take the thermal conductivity of cupro-nickel alloys from Table 1, 50

W/m°C, the fouling coefficients from Table 3; methanol (light organic)
5000 Wm-2°C-1, brackish water (sea water), take as highest value, 3000
Wm-2°C-1 do
do ln
1 1 1 di do 1 do 1
= + + + × + ×
Uo ho hod 2K w di hid di hi
-3 20
1 1 1 20 × 10 ln 20 1 20 1
= + + 16 + × + ×
Uo 2740 5000 2 × 50 16 3000 16 3812

Uo = 738 W m2  C

Well above assumed value of 600 Wm-2°C

42
 References

1. EAGLE, A. and FERGUSON, R. M. (1930) Proc. Roy. Soc.

A. 127, 540. On the coefficient of heat transfer fromthe
internal surfaces of tube walls.
2. KERN, D. Q. (1950) Process Heat Transfer (McGraw-Hill).
3. LUDWIG, E. E. (2001) Applied Process Design for
Chemical and Petroleum Plants, Vol. 3, 3rd edn (Gulf).
4. PERRY, R. H., GREEN, D.W. and MALONEY, J. O. (1997)
Perry’s Chemical Engineers Handbook, 7th edn (McGraw-
Hill).
5. SAUNDERS, E. A. D. (1988) Heat Exchangers
(Longmans).
43
2/1/2009 8:11:44 AM 44
 Step 5
Table. 3: Fouling factors (coefficients), typical values

45