You are on page 1of 37

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/339127185

HEAT EXCHANGER DESIGN REPORT DESIGN OF A SHELL AND TUBE HEAT


EXCHANGER, 1 SHELL -1 TUBE PASS WITH FIXED TUBE HEAD AND COUNTER
CURRENT FLOW STATEMENT OF AUTHORSHIP Heat Exchanger De...

Technical Report · February 2020


DOI: 10.13140/RG.2.2.25325.13285

CITATIONS READS

0 8,124

1 author:

Phelelani Mamba
University of KwaZulu-Natal
1 PUBLICATION   0 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Heat exchanger design View project

All content following this page was uploaded by Phelelani Mamba on 08 February 2020.

The user has requested enhancement of the downloaded file.


HEAT EXCHANGER DESIGN REPORT

DESIGN OF A SHELL AND TUBE HEAT EXCHANGER, 1 SHELL - 1


TUBE PASS WITH FIXED TUBE HEAD AND COUNTER CURRENT
FLOW

By

Mamba Phelelani Eshmael

Student number: 217058111

Group Number: Four

Date: 16 October 2019

Design Unit: Heat Exchanger E-607

TEMA Type: AEL

Supervisor(s): Mrs A. Singh & Dr K. Osman

Submitted in partial fulfilment of Academic Requirements for the course ENCH3ED in the
curriculum for the degree of Bachelor of Science in Chemical Engineering at the University
of KwaZulu-Natal
STATEMENT OF AUTHORSHIP

Heat Exchanger Design

by

Phelelani Eshmael Mamba

I hereby declare that this design and the associated report is my own work
(except where formally acknowledged in the section headed "acknowledgement")*

Student Number: 217058111 Signature: _________________________

Date: 14 October 2019

1
Abstract
In this present work, a shell and tube heat exchanger with a 1-shell, 1-tube pass and counter-
current flow designed from Kern’s method is presented. The rear-end type was specified to be
a fixed tube sheet. The TEMA type was selected to be AEL. Heat exchanger labelled E-607 on
a cumene production plant was designed to cool cumene-diisopropyl benzene (DIPB) mixture
from 80 to 25 °C. The mass flow rate of the cumene-DIPB mixture flowing on the tube side
was 3.83 kg/s with an over-design factor of 10 % incorporated to accommodate increase in
production rate. Initially, polyalphaolefin (PAO) was the proposed cooling fluid. However,
PAO resulted in high-pressure drops on the shell and tube side due to its relatively high
viscosity. Dowtherm SR-1 fluid composed of each 50 % ethylene glycol and water by volume
was used as an alternative cooling fluid flowing on the shell-side at 4.60 kg/s. The inlet and
outlet temperatures of Dowtherm SR-1 were 5.00 °C and 31.23 °C respectively, and the log
mean temperature difference across the unit was 32.28 °C. The heat duty required to attain the
above-stated cooling was 395.73 kW. All fluid physical properties were evaluated at mean
temperatures due to relatively small changes in temperatures of both streams across the unit.
ASME/ANSI carbon steel pipes of ½ inch nominal bore, schedule XXS, with an external
diameter of 21.336 mm, internal diameter of 6.4 mm and thickness of 7.468 mm were used in
the fabrication of the unit. A triangular arrangement of tubes with a pitch of 26.67 mm was also
used. From an estimated overall heat transfer coefficient of 430 W/m2 k, a total required heat
transfer area of 28.51 m2 was determined. The calculated overall heat transfer coefficient was
331.10 W/m2 k, which was 23.00 % less than the estimated value and the difference was within
specification. 90 tubes of length 4.88 m were used to make up the heat transfer area of 28.51
m2. The tube bundle diameter was 297.25 mm, and a clearance length of 12.00 mm was added
to obtain an inside shell diameter of 309.25 mm. A 4.88 m length ASME/ANSI carbon steel
pipe of 14-inch nominal bore, schedule 80, with an external diameter of 355.60 mm, internal
diameter of 317.50 mm and thickness of 19.05 mm was selected for making up shell of the
unit. The velocity of the tube and shell side fluid was 1.58 m/s and 0.519 m/s, respectively and
were both within specification. The shell side consisted of 25 % horizontal cut segmental
baffles with a spacing of 0.130 m apart. Kern’s method was used to determine the pressure
drop on the shell side. The pressure drop on the shell and tube side was 55.19 kPa and 31.67
kPa respectively. The total cost of fabricating the unit was obtained to be R 291 238.501 .

1
The cost of the heat exchanger is correct at the time of publishing this work

2
Table of Contents
1. Introduction ............................................................................................................................ 1

1.1 Report Objectives............................................................................................................ 1

1.2 Project statement ............................................................................................................. 1

1.3 Scope of the project ......................................................................................................... 2

2. Theoretical background ......................................................................................................... 2

Method of Approach .............................................................................................................. 2

2.1 Thermal and mechanical design calculations .................................................................. 3

2.1.1 The Energy Balance .................................................................................................. 3

2.1.2 The Log Mean Temperature Difference and Heat Transfer Area ............................. 4

2.1.4 Number of Tubes and Pitch ...................................................................................... 4

2.1.4 Tube Cross-Section Area and Tube-Side Velocity ................................................... 5

2.1.5 Bundle Diameter and Shell Diameter ....................................................................... 5

2.1.6 Tube Side Heat Transfer Coefficient ........................................................................ 6

2.1.7 The Shell Side Velocity ............................................................................................ 6

2.1.8 Shell Side Heat Transfer Coefficient ........................................................................ 7

2.1.7 The Overall Heat Transfer Coefficient ..................................................................... 7

2.1.8 Tube Side Pressure Drop .......................................................................................... 7

2.1.10 Shell Side Pressure Drop ........................................................................................ 8

Material Selection .............................................................................................................. 8

Costing ............................................................................................................................... 8

Results ........................................................................................................................................ 9

Discussions .............................................................................................................................. 10

Conclusions .............................................................................................................................. 17

Recommendations .................................................................................................................... 18

Acknowledgements .................................................................................................................. 18

Appendices ............................................................................................................................... 19

3
Appendix A: Nomenclature ................................................................................................. 19

Appendix B .......................................................................................................................... 20

References ............................................................................................................................ 20

Appendix C: Piping and Instrumentation diagrams ............................................................. 21

Appendix D: Figures and tables of sourced data ................................................................. 22

Appendix E: Fluids physical properties and sample calculations ........................................ 24

List of Tables
Table 1: Thermal and mechanical design specifications of heat exchanger E-607 ................... 9
Table 2: Physical properties of PAO and Dowtherm at a mean temperature of 20 °C (Ghajar &
Tang, 1994) and (DOW chemical company, 2011) ................................................................. 11
Table 3: Comparison of design parameters for PAO and Dowtherm cooled STHE ............... 15
Table E 1: Physical Properties of pure components retrieved using the NRTL model in Aspen®
Plus Version 10.0 and (DOW chemical company, 2011) ........................................................ 25
Table E 2: Physical properties of fluids at wall temperature retrieved using the NRTL model in
Aspen® Plus Version 10.0 and (DOW chemical company, 2011) .......................................... 25
Table E 3: Bundle diameter coefficients for triangular pitch .................................................. 25
Table E 1: Physical Properties of pure components retrieved using the NRTL model in Aspen®
Plus Version 10.0 and (DOW chemical company, 2011) ........................................................ 25
Table E 2: Physical properties of fluids at wall temperature retrieved using the NRTL model in
Aspen® Plus Version 10.0 and (DOW chemical company, 2011) .......................................... 25
Table E 3: Bundle diameter coefficients for triangular pitch (Sinnott, 2005) ......................... 25
List of Figures
Figure C 1: Section of the process flow diagram near heat exchanger E-607 ......................... 21
Figure C 2: Heat exchanger mechanical drawing with and dimensions in mm ...................... 21
Figure C 3: Cross sectional view of heat exchanger tube layout with dimensions in millimeters
.................................................................................................................................................. 22
Figure D 1: Tube-Side friction factors (Sinnott, 2005)............................................................ 22
Figure D 2: Tube-Side heat transfer factor (Sinnott, 2005) ..................................................... 22
Figure D 3: Shell and tube heat exchanger purchase cost for mid-2004 ................................. 23
Figure D 4: Shell-Side heat transfer factors (Sinnott, 2005) ................................................... 23

4
Figure D 5: Shell-Side friction factors (Sinnott, 2005)............................................................ 23
Figure D 6: Bundle to Shell diameter clearance length (Sinnott, 2005) .................................. 24
Figure E 1: Temperature profiles for shell and tube side fluid ................................................ 24

Figure D 1: Tube-Side friction factors (Sinnott, 2005)............................................................ 27


Figure D 2: Tube-Side heat transfer factor (Sinnott, 2005) ..................................................... 27
Figure D 3: Shell and tube heat exchanger purchase cost for mid-2004 ................................. 28
Figure D 4: Shell-Side heat transfer factors (Sinnott, 2005) ................................................... 28
Figure D 5: Shell-Side friction factors (Sinnott, 2005)............................................................ 28
Figure D 6: Bundle to Shell diameter clearance length (Sinnott, 2005) .................................. 29

1. Introduction

Heat transfer units are of paramount importance in regulating stream temperatures within
chemical industries. The geometric design of a heat exchanger can be of any type including
shell and tube heat exchangers, double pipes and plate and frame heat exchangers. Depending
on the heat duty required to achieve the desired temperature of a stream, the heat exchanger
type can be selected accordingly. Shell and tube heat exchangers (STHE) are widely used in
industries because of their large heat transfer area over a small floor space (Sinnott, 2005).
Moreover, the well-known rigorous design procedures available in the open literature such as
Kern’s and Bell’s method used for the mechanical and thermal design of these units result in
efficient and optimum designs.

1.1 Report Objectives


➢ Produce detailed thermal design specifications of heat exchanger E-607 based on a
cumene production plant.
➢ Specification of the unit using TEMA codes (i.e. BEM, AEM, NEN)
➢ A clear and concise justification of fluid allocation within the unit.
➢ Selection of material for construction and dimensions of the unit including but not
limited to the tube length, shell and tube diameters.
➢ Presentation of the Process and instrumentation diagram (PID) and a detailed cross-
sectional drawing of the unit.
➢ Rigorous cost estimations for the initial fabrication of the unit

5
1.2 Project statement
In this present work, the design of a shell and tube heat exchanger unit was carried out using
design specifications provided by Process Design Associates (Ltd). The task at hand was to
design a shell and tube heat exchanger to cool a cumene-diisopropyl benzene mixture rich in
cumene emerging as a distillate product from a distillation unit. The heat exchanger was based
on a cumene production plant depicted in figure C 1 of appendix C. The specifications of heat
exchanger unit labelled E-607 on the above-mentioned figure were, 1-shell, 1-tube pass with
fixed tube sheets on the front and rear end. The flow pattern was also specified to be counter-
current flow, and polyalphaolefin (PAO) was the proposed cooling fluid. However, due to high
pressure drops experienced on the shell and tube side, Dowtherm SR-1 fluid was used as a
substitute for PAO.

1.3 Scope of the project


The thermal design of the unit had to adhere to international standards stipulated by Tubular
Exchanger Manufacturers Association, Inc. (TEMA). Moreover, the material of construction,
dimensions of tubes and shell sizes had to conform to international standards stipulated by the
American Society of Mechanical Engineers (ASME).

2. Theoretical background

Method of Approach
Structure of calculations for the design of heat exchanger E-607.

➢ Analysis of process specification, including temperatures, flow rates, and compositions


and familiarization with operational limits, including permissible pressure drops,
velocity ranges, mechanical ratios, and standard lengths.
➢ Determination of the heat duty across the unit using a flow rate with a 10% design
safety factor incorporated.
➢ Collection of stream properties, namely density, dynamic viscosity, heat capacities, and
thermal conductivities at mean temperatures of both streams.
➢ Estimation of the overall heat transfer coefficient for the heat exchanger.
➢ Calculation of the log mean temperature difference across the unit.
➢ Determination of the heat transfer area.
➢ Fluid allocation to the tube and shell side and tube layout.
➢ Calculation of the total number of tubes required and the tube-side velocity.
➢ Determination of tube bundle and shell diameter.

6
➢ Calculation of the tube-side heat transfer coefficient.
➢ Selecting and calculating baffle type, baffle spacing, equivalent shell diameter, shell-
side velocity, and shell-side heat transfer coefficient.
➢ Determination of Fouling factors and calculation of the overall heat transfer coefficient.
➢ Calculation of the tube-side pressure drop.
➢ Calculation of the shell-side pressure drop using Kern’s Method
➢ Initial costing of the entire project.

Shell and tube heat exchangers are usually made from a bundle of tubes that are held together
by tube sheets on the front and rear end, respectively. The tube sheet is also useful in preventing
mixing of the shell and tube side fluids near ends of the unit (Mukherjee, 1998). Fixed tube
sheet heat exchangers are generally regarded to be less costly more especially when no
expansion joint on the rear end is included (Mukherjee, 1998). Moreover, mechanical cleaning
on the tube side is possible upon the removal of the cover heads on both ends. However, the
shell side is not accessible for mechanical cleaning due to tube sheets welded on the shell hence
fixed tube sheets heat exchangers are not suitable for heavily fouling fluids on the shell side
unless cleaning by chemical means is effective (Mukherjee, 1998).

Baffles are also included along the length of the tube bundle. They serve to hold the tubes in
place and to create a turbulent flow regime on the shell side hence increasing the shell side heat
transfer coefficient and amount of heat transferred (Sinnott, 2005). The percentage cut and
spacing of baffle cuts should be carefully selected to prevent high-pressure drops across the
unit. High-pressure drops may induce mechanical vibrations that can reduce the lifespan of the
unit and increase the annual pumping costs (Serth, 2007). Optimum baffle cut of 20 – 30 %
and baffle spacings of (0.3 – 0.5) ratio of the outer shell diameter are recommended (Sinnott,
2005).

2.1 Thermal and mechanical design calculations


Kern’s method will be adopted in the design of the unit due to its satisfactory predictions of
the preliminary specifications of shell and tube heat exchangers. Moreover, because the method
was developed based on experimental data, it is mostly recommended in the literature for
producing optimum designs (Sinnott, 2005).

7
2.1.1 The Energy Balance
For purposes of this design, steady-state conditions will be assumed across the unit. Fluid
properties, including the specific heat capacity, viscosity and density will be assumed to be
constant and temperature independent. The unit will also be assumed to have negligible heat
loss to the surroundings, and the fluids will also be assumed to have negligible potential and
kinetic energy due to minimum elevations and low velocities across the unit, respectively.
Moreover, the Log Mean Temperature Difference approach (LMTD) will be adopted. From
the assumptions laid above equation (E-2.1.1) can be used to determine the heat duty across a
heat exchanger.

𝑄 = 𝑚̇𝐶𝑃 𝑑𝑇 (E – 2.1.1)

The above relationship can be used for both the cold and hot stream depending on which stream
is fully specified. From the negligible heat loss assumption, equation 2.1.2 is valid.

𝑄𝑐𝑜𝑙𝑑 = 𝑄ℎ𝑜𝑡 (E – 2.1.2)

2.1.2 The Log Mean Temperature Difference and Heat Transfer Area
The log mean temperature difference (𝑇𝑙𝑚 ) (LMTD) is defined as the true average temperature
difference across the tube length as it accounts for the exponential decay nature of temperature
along the tube (Lavine, et al., 2011). Using the linear form of the temperature difference
between the inlet and outlet of tubes may over-estimate the heat duty, thus leading to poor
designs.

(𝑇1 −𝑡2 )−(𝑇2 −𝑡1 )


𝛥𝑇𝑙𝑚 = 𝑇 −𝑡 (E – 2.1.3)
𝑙𝑛( 1 2)
𝑇2 −𝑡1

𝑄 = 𝐴0 𝑈0 𝛥𝑇𝑙𝑚 (E – 2.1.4)

Depending on the number of shell and tube passes, the LMTD is usually corrected by
incorporating a correction factor (𝐹𝑡 ) on the heat duty equation presented in (E – 2.1.4). The
modified heat duty equation (E – 2.1.5) is then adopted in calculating the heat duty across a
heat exchanger. The azimuthal heat transfer area is the effective area where heat transfer is
taking place. The overall heat transfer coefficient can be assumed based on estimated literature
values.

𝑄 = 𝐴0 𝑈0 𝐹𝑡 𝛥𝑇𝑚 (E – 2.1.5)

Rearranging for the heat transfer area 𝐴0 , equation 2.1.6 is obtained.

8
𝑄
𝐴0 = 𝑈 (E – 2.1.6)
0 .𝐹𝑡 .𝛥𝑇𝑚

2.1.4 Number of Tubes and Pitch


The choice of tube sizes selected for the design of a heat exchanger has a significant effect on
both the mechanical and thermal design of the unit. Pipe diameters in the range of (3/8 to 1)
inch nominal diameter are generally used for the design of heat exchangers (Sinnott, 2005). In
this range, the ¾ inch OD pipe sizes are commonly used (Kern, 1950). Pipes sizes less than ¾
inch OD should not be used for heavily fouling fluids due to easy blockage of pipes which may
increase maintenance costs (Mukherjee, 1998). However, smaller tube sizes are ideal in
increasing the tube side heat transfer coefficient; hence, the overall heat transfer coefficient due
to increased velocities through the pipe. Pipes sizes less than ¾ inch OD can only be used in
smaller heat exchangers with heat transfer areas less than 20 – 30 m2 (Mukherjee, 1998). The
tube length can be selected from standard pipe lengths stipulated by ASME. Once the tube
length and diameter have been selected, the total number of tubes (𝑁𝑡 ) can be determined.

𝐴𝑡𝑢𝑏𝑒 = 𝜋𝑑0 𝐿 (E – 2.1.7)

𝐴0
𝑁𝑡 = 𝐴 (E – 2.1.8)
𝑡𝑢𝑏𝑒

2.1.4 Tube Cross-Section Area and Tube-Side Velocity


From the selected tube size, the inner pipe diameter is used to determine the cross-section area
(𝐴𝑐 ) for one tube.

𝜋(𝑑𝑖 )2
𝐴𝑐 = (E – 2.1.9)
4

From the tube side mass flow rate and fluid density (𝜌𝑡 ), the volumetric flow rate and hence
the velocity of the fluid through the circular conduits can be quantified.

𝑚̇
𝑄̇𝑡 = 𝜌 (E – 2.1.10)
𝑡

𝑄̇𝑡
𝑢𝑡 = 𝐴 (E – 2.1.11)
𝑡𝑢𝑏𝑒

(Sinnott, 2005) Recommended tube-side velocities in the range of 1 – 2 m/s to prevent


suspended solids from settling on the tube walls and cause fouling in the unit.

9
2.1.5 Bundle Diameter and Shell Diameter
The bundle diameter (𝐷𝑏 ) made up by the total number of tubes and the pitch (spacing) between
tubes is used to determine the shell diameter (𝐷𝑠 ).
1
𝑁 𝑛
𝐷𝑏 = (𝐾𝑡 ) 1 (E – 2.1.12)
1

𝐷𝑠 = 𝐷𝐵 + 𝑐𝑙𝑒𝑎𝑟𝑎𝑛𝑐𝑒 𝑙𝑒𝑛𝑔𝑡ℎ (E – 2.1.13)

For a triangular tube arrangement, the tube pitch can be determined as shown below.

𝑃𝑡 = 1.25 𝑑𝑜 (E – 2.1.14)

Values of 𝑛1 and 𝐾1 are obtained from table E 3 in appendix E. Moreover, the clearance length
can also be obtained from figure D 6 in appendix D.

2.1.6 Tube Side Heat Transfer Coefficient


The tube side heat transfer coefficient can be estimated from the correlation valid for less
viscous fluids in turbulent flow regime.

ℎ𝑖 𝑑𝑖 𝜇 0⋅14
𝑁𝑢 = = 𝑗ℎ . 𝑅𝑒. 𝑃𝑟0.33 (𝜇 ) (E – 2.1.15)
𝑘𝑓 𝑤

Rearranging for the tube side heat transfer coefficient (ℎ𝑖 ) and other parameters :

𝑁𝑢 .𝑘𝑓
ℎ𝑖 = (E – 2.1.16)
𝑑𝑖

𝐶𝑃 .𝜇
𝑃𝑟 = (E – 2.1.18)
𝑘𝑓

𝜇 0⋅14
The viscosity correction factor (𝜇 ) usually deviates from unity for highly viscous fluids
𝑤

(Sinnott, 2005). The relationship below can be used to estimate the tube wall temperature (𝑡𝑤 )
and the viscosity of the fluid at the wall temperature can be determined.

ℎ𝑖 (𝑡𝑤 − 𝑡) = 𝑈𝑜𝑣𝑒𝑟𝑎𝑙𝑙 (𝑇 − 𝑡) (E – 2.1.19)

Where 𝑡𝑤 , 𝑡 and 𝑇 is the estimated wall temperature, tube-side bulk temperature and shell-side
bulk temperature, respectively. Rearranging for 𝑡𝑤 , E – 2.1.20 is obtained.

𝑈𝑜 (𝑇−𝑡)+ h𝑖 .𝑡
𝑡𝑤 = (E – 2.1.20)
ℎ𝑖

10
The mixture kinematic viscosity (𝜈) and hence, the dynamic viscosity (𝜇) can be estimated
from the Gambill method (Gambill, 1959).

1⁄ 1⁄ 1⁄
3 3 3
𝜈𝑚𝑖𝑥𝑡𝑢𝑟𝑒 = 𝑥𝑎 𝜈𝑎 + 𝑥𝑏 𝜈𝑏 (E – 2.1.21)

Other mixture physical properties, including the thermal conductivity, density and specific heat
capacity are obtainable from equation 2.1.22 below.

∑𝑁
𝑖=1 𝑥𝑖 𝜃𝑖 (E – 2.1.22)

2.1.7 The Shell Side Velocity


The Area of cross-flow (𝐴𝑠 ) can be determined from the relationship below.

(𝑃𝑡 −𝑑𝑜 )𝐷𝑠 .𝑙𝑏


𝐴𝑠 = (E – 2.1.23)
𝑃𝑡

From the area of cross-flow, the shell side velocity (𝑢𝑠 ) can be determined as follows.

𝑚̇𝑡
𝐺𝑠 = (E – 2.1.24)
𝐴𝑠

𝐺
𝑢𝑠 = 𝜌𝑠 (E – 2.1.25)
𝑠

The shell side velocity is recommended to be in the range of 0.3 – 1.0 m/s for minimum
pressure drop.

2.1.8 Shell Side Heat Transfer Coefficient


A fluid resulting in high heat transfer coefficients is desired. (Mukherjee, 1998) Postulated
two relationships relating the heat transfer coefficient and physical properties; mainly, ℎ ∝
(𝜇)−0⋅47 and ℎ ∝ (𝑘𝐹 )0⋅67 . The two relations indicate that a fluid with low viscosity
and high thermal conductivity should be adopted to attain high heat transfer coefficients. The
shell side equivalent diameter (𝑑𝑒 ) for triangular tube arrangement is determined as shown.

1.11
𝑑𝑒 = ( 𝑑 ) . (𝑃𝑡2 − 0.917𝑑𝑜2 ) (E – 2.1.26)
0

The Reynolds number and hence the shell side heat transfer coefficient on the shell side is
calculated using the equivalent diameter as shown below.

𝜌.𝑢𝑠 .𝑑𝑒
𝑅𝑒 = (E – 2.1.27)
𝜇

0⋅14
ℎ 0 𝑑𝑒 𝜇
𝑁𝑢 = 𝑘𝑓
= 𝑗ℎ . 𝑅𝑒 . 𝑃0⋅33
𝑟 (𝜇 ) (E – 2.1.28)
𝑤

11
The 𝑗ℎ heat transfer factor can be obtained by interpolation from figure D 4 of appendix D.

𝑁𝑢 𝑘𝑓
ℎ0 = (E – 2.1.29)
𝑑𝑒

2.1.7 The Overall Heat Transfer Coefficient


To account for fouling that reduces the efficiency of a heat exchanger over time, the overall
heat transfer coefficient calculated should incorporate the fouling resistances of both fluids
flowing on the shell and tube side, respectively.

𝑑
1 1 𝑑0 × ln( 0 ) 1 𝑑
𝑑𝑖
= ℎ + 𝑅𝑓, + + (ℎ + 𝑅𝑓, 𝑖 ) × ( 𝑑0 ) (E – 2.1.30)
𝑈0 0 2 × 𝑘𝑤 𝑖 𝑖

2.1.8 Tube Side Pressure Drop


The pressure drop on the tube side can be determined from (E – 2.1.31) below.

𝐿 𝜇 −𝑚 𝜌𝑢𝑡
𝛥𝑃𝑡 = 𝑁𝑃 (8 × 𝑗𝑓 × (𝑑 ) × (𝜇 ) + 2.5 ) × 2
(E – 2.1.31)
𝑖 𝑤

Where m = 0.25 for laminar flow (𝑅𝑒 < 2100) and m = 0.14 for turbulent flow (𝑅𝑒 > 2100).
The 𝑗𝑓 friction factor is obtained by interpolation from figure D 1 of appendix D.

2.1.10 Shell Side Pressure Drop


Kern’s method can be used to determine the shell side pressure drop, as shown below.

−0⋅14
𝐷 𝐷 𝜌𝑢2𝑠 𝜇
𝛥𝑃𝑠 = 8 × 𝑗𝑓 × ( 𝑑𝑠 ) × ( 𝑑𝑠 ) × (
2 𝜇𝑤
) (E – 2.1.32)
𝑒 𝑒

(Sinnott, 2005) Recommended that the pressure must be below 35 kPa for the fluids with a
viscosity less than 1 mPa-s and must be in the range of 50 – 70 kPa for fluids with a viscosity
in the range of 1 – 10 mPa-s.

Material Selection
The choice of material for construction in heat exchangers is usually influenced by the process
fluids and operating conditions such as temperature and pressure. An ideal material for
construction is desired to have low corrosion rate, high tensile strength, high thermal
conductivity, less cost and high reliability (Lancaster, 1983). From the spectrum of steels
available in the market, (Lancaster, 1983) stated that carbon steel and Austenitic chromium-
nickel steel are the most used materials for general purpose heat exchangers. In the temperature
range of 0 – 500 °C, both materials give satisfactory results, and for temperatures above 500
°C, refractory-lined Austenitic steel is typically employed (Lancaster, 1983).

12
Carbon steel is mostly considered for less corrosive fluids such as hydrocarbons while
Austenitic steel may also be considered for use in mildly corrosive fluids such mercaptan
hydrocarbons mostly found in the oil and gas industries (Lancaster, 1983). In general,
Austenitic steel is classified as high-grade steel, thus more costly and reliable compared to
carbon steel. Moreover, due to the presence of chromium in Austenitic steel, corrosion effects
are reduced significantly. To alleviate the effect of corrosion in materials such as carbon steel,
coatings such as epoxy phenolic resins are usually applied on surfaces for protection against
corrosion (Lancaster, 1983).

Costing
An initial estimate of the purchasing price of the heat exchanger can be done by using the factorial
method outlined by (Sinnott, 2005). By using figure D 3 in appendix D and accounting for inflation
and necessary currency conversion rates, the cost of the unit can be obtained.

13
Results
Table 1: Thermal and mechanical design specifications of heat exchanger E-607

Remember to open the design suggestion document

14
Discussions

From design specifications stipulated by Process Design Associates (Ltd), the design of heat
exchanger E-607 based on a cumene production plant shown in figure C 1 of appendix C was
carried out. The unit was specified to be a 1-shell, 1-tube pass with fixed tube sheets on the
front and rear end, respectively. Counter-current flow arrangement was also specified, and
Polyalphaolefin (PAO) fluid was initially proposed as a coolant in the unit. The unit was
mainly designed to cool the liquid distillate product (cumene-DIPB mixture) emanating from
column D-604 from 80 °C to 25 °C.

The cooling shell-side fluid was assumed to be available at a pressure of 1 atm from its
reservoir, and the hot tube-side fluid from the column was specified to be at 1 atm. From these
pressure specifications, the heat exchanger TEMA designation was selected to be AEL. The A
front cover which is a channel with a removable cover is suitable for use in low-pressure tube
side fluids in heat exchangers (Serth, 2007). Since the pressure of the tube side fluid was
relatively low, the A front cover was therefore suitable for use in this design. Moreover,
because the A-type front cover can be easily detached from the unit, access to the tubes for
mechanical cleaning is made simple. The shell of the unit denoted by E on the TEMA code
was trivial from the specification of 1-shell pass. The Rear end denoted by L was selected
based on justifications made for the front-end cover. Moreover, having similar covers on both
ends allows access to tubes for cleaning purposes possible on both ends of the heat exchanger.

In this work, all fluid physical properties were retrieved at mean temperatures of each
respective stream because of relatively small temperature changes across the heat exchanger.
From the design project statement, polyalphaolefin (PAO) was the proposed fluid to be used in
cooling the mixture2. However, polyalphaolefin (PAO) has a specific heat capacity of 2126.2
J/kg K at a mean temperature of 20 °C which is comparable to that of the mixture which is
1878.76 J/kg K at its mean temperature of 52.5 °C. The close range in specific heat capacities
of the two fluids causes PAO on the shell side to exit at a relatively high temperature
approaching the inlet temperature of the hot stream. This causes a decrease in the log mean
temperature difference hence a high heat transfer area in the heat exchanger. To achieve the
required heat transfer area, longer tubes are required, and the number of tube passes need to be
increased, which increases the overall cost and fabrication complexity of the equipment.
Moreover, due to the high viscosity of PAO, significant pressure drops beyond the maximum

2
The term “mixture” refers to the cumene-DIPB mixture (may be used interchangeably for convenience)

15
allowable pressure drop were experienced on the shell side. Allocating PAO on the tube side
was not permissible since it is more viscous than the mixture.

To alleviate the above-mentioned effects, Dowtherm SR-1 fluid composed of each 50 % water
and ethylene glycol was used as a substitute for PAO. From the two mathematical relationships
postulated by (Mukherjee, 1998) that relates the heat transfer coefficient with the viscosity and
thermal conductivity of a fluid, a highly viscous fluids with low thermal conductivity result in
low heat transfer coefficients hence a high heat transfer area. From the comparison of physical
properties of PAO and Dowtherm in table 2, Dowtherm has a higher thermal conductivity and
low viscosity when compared to PAO fluid. Therefore, Dowtherm physical properties promote
high heat transfer coefficients and hence less heat transfer area in the heat exchanger.
Moreover, the specific heat capacity of Dowtherm is relatively higher than that of the mixture.
This aid in increasing the log mean temperature difference across the unit. The relationships
postulated by (Mukherjee, 1998) are repeated here for convenience.

ℎ ∝ (𝜇)−0⋅47 and ℎ ∝ (𝑘𝐹 )0⋅67

Table 2: Physical properties of PAO and Dowtherm at a mean temperature of 20 °C (Ghajar & Tang,
1994) and (DOW chemical company, 2011)

Physical property Polyalphaolefin Dowtherm SR-1


Viscosity (mPa-s) 7.09 3.94
Thermal conductivity (W/m2 K) 0.137 0.367
Specific heat capacity (J/mol K) 2126.2 3282

In addition, Dowtherm SR-1 fluid was also selected as a substitute based on its exceptional
properties, including low toxicity, low flammability and minimum corrosion effects on metals
due to the presence of corrosion inhibitors (DOW chemical company, 2010)

The cumene-DIPB mixture was allocated on the tube side due to its relatively low viscosity
compared to that of Dowtherm. Allocating viscous fluids on the tube side results in high-
pressure drops due to relatively small diameter tubes typically used in heat exchangers. High-
pressure drops cause mechanical induced vibrations which may reduce the life span of the unit.
Moreover, since the mixture was the hotter fluid, Kern’s method suggests that hot fluids should
be allocated on the tube side for optimum performance of a heat exchanger. Both fluids had
the same fouling factors; however, Dowtherm SR-1 fouling deposits can be adequately cleaned
by chemical means such as using inhibited acids followed by neutralisation and phosphatisation
(DOW chemical company, 2010). To allow mechanical cleaning of deposits from the mixture,

16
tube side allocation was preferable since mechanical cleaning on the shell side is impossible in
a fixed tube sheet shell and tube heat exchanger. The presence of corrosion inhibitors in
Dowtherm SR-1 fluid possibly makes it less corrosive compared to the mixture, thus allowing
it to be allocated on the shell side.

The flow rate of the mixture flowing on the tube side was initially specified to be 3.48 kg/s.
However, to account for operational upset within the plant which might be inherent to increased
production rates, a 10 % design factor was incorporated thus resulting in a total tube side flow
rate of 3.83 kg/s. Moreover, the design factor included here serves to give flexibility in process
operation and account for any possible contingencies within the process. On the other hand,
the flowrate of Dowtherm on the shell was set to 4.60 kg/s, which is 20 % greater than the tube
side flowrate. Having the flowrates of the tube and shell side within a 20 % difference was one
of the specifications in the design of the unit. Shell side (Dowtherm) inlet temperature was
selected to be 5 °C, while the outlet was determined to be 31.23 °C. The inlet temperature of 5
°C is relatively low, and it may require external cooling of Dowtherm before being introduced
to the heat exchanger. However, this was the maximum allowable inlet temperature in order to
attain a temperature approach of 20 °C between the tube side outlet and shell side inlet fluid.
As recommended by (Sinnott, 2005), a high-temperature approach in a heat exchanger is ideal
in reducing the heat transfer area. If the provision of Dowtherm at 5 °C appears to be a poor
design choice, comparing the cost of using a refrigerant as a coolant may be considered. The
high shell side flowrate was used in order to reduce the outlet temperature of the shell side
fluid. Lower exit temperatures on the shell side increase the log mean temperature difference
in heat exchangers. The log mean temperature difference across the unit was determined to be
32.28 °C.

From an assumed overall heat transfer coefficient of 430 W/m2K and a total heat duty of 395.73
kW, a total heat transfer area of 28.51 m2 across the unit was determined. Tubes to fabricate
the unit were selected from a spectrum of AMSE standard pipe sizes. The use of standard pipe
sizes reduces the design freedom in heat exchangers. However, the cost of customising tubes
from manufacturers is eliminated hence reducing the total cost of fabricating heat exchangers.
ASME/ANSI carbon steel pipes of ½ inch nominal bore, schedule XXS, with an external
diameter of 21.336 mm, internal diameter of 6.4 mm and thickness of 7.468 mm were selected
for use in fabricating the unit. The triangular arrangement of tubes was selected so to attain
high heat transfer coefficients within the heat exchanger. From the selected tube diameters, 90
tubes of length 4.88 m were needed to make up the required heat transfer area. To prevent

17
settling of suspended particles causing fouling in heat exchangers, the velocity of the mixture
through the tubes was set to 1.58 m/s which is in the range of 1 – 3 m/s recommended by
(Sinnott, 2005).

Since the Cumene-DIPB mixture is a mercaptan free hydrocarbon at a temperature below 500
°C, its corrosivity is considered to be bearable by carbon steel hence making carbon steel the
material of choice as recommended by (Lancaster, 1983). Corrosion inhibitors present in
Dowtherm reduces the corrosion rate in carbon steel, thus making carbon steel compatible with
both the tube and shell side fluid. Moreover, to extend the lifespan of carbon steel and reduce
the possibility of tube perforation, coating the tubes using epoxy phenolic resins will be
implemented as suggested by (Lancaster, 1983). Even though high-grade steel such as
austenitic steel recommended by (Lancaster, 1983) do not need the application of coatings to
withstand corrosion because of their unique metallurgical properties, their purchase is
expensive compared to that of carbon steel.

For small flowrates through heat exchangers, smaller tube diameters are required to achieve
high velocities unless the number of tube passes is increased. Increasing the number of tube
passes has the disadvantage of increasing the tube side pressure drop and fabrication cost of
heat exchangers. To extend this discussion, (Mukherjee, 1998) stated that tubes of diameters
less than ¾ inch might limit mechanical cleaning, and they are more prone to blockage. In this
design, the stringent specification on the number of tube passes was stated to be one tube pass,
and tube side mass flow rate was fixed at 3.83 kg/s. Therefore, the use of smaller tube diameters
less than ¾ inch was inevitable in order to attain a tube-side velocity in the range of 1 – 2.0
m/s. Having the heat transfer area of 28.51 m2 which is in the range of 20 - 30 m2 classifies
the unit as a small heat exchanger according to (Mukherjee, 1998). (Mukherjee, 1998) Stated
that the use of small diameter tubes below ¾ inch is permissible for small heat exchangers
having a heat transfer area in the range of 20 – 30 m2.

Selecting tubes for constructing a heat exchanger requires careful consideration on the tube
wall thickness. (Kern, 1950) Stated that tubes of a wall thickness greater than 3.2 mm should
be used in heat exchangers to give a corrosion allowance of 3.2 mm and prevent perforation of
tubes. The use of 7.468 mm tubes in this design satisfied the recommendation stated by (Kern,
1950). As a pragmatic approach in designing a heat exchanger, accounting for the tube sheet
thickness on both ends of the unit is a necessity. The tube sheets were made from carbon steel
to prevent galvanic corrosion caused by the difference in electric potential between two

18
different metals that are in contact in the presence of a conducting fluid. The tube sheet
thickness was assumed to be 25 mm each, and the total effective heat transfer area was
calculated from a tube length of 4.83 m. In theory, the above-described calculation assumes
that the length covered by tube sheets on both ends do not contribute to heat transfer between
the shell and tube side fluid.

As stated earlier in this work, the selected triangular arrangement of tubes promotes high heat
transfer coefficients in heat exchangers. However, such arrangement resulted in a high-pressure
drop of PAO on the shell side. The use of a less viscous fluid such as Dowtherm sufficed in
circumventing the issue of high-pressure drops experienced on the shell side. The attempt of
using a square arrangement of tubes resulted in a low shell side heat transfer coefficient, which
also resulted in a low overall heat transfer coefficient across the heat exchanger. Even though
the square arrangement of tubes permits easy cleaning of tubes on the outside, such an
advantage cannot be exploited in a fixed tube sheet shell and tube heat exchanger as the shell
side is not accessible.

The bundle diameter made up by 90 tubes was determined to be 0.2972 m (297.2 mm). An
inner shell diameter of 0.30925 m (309.25 mm) was obtained by adding a 12 mm clearance on
the bundle diameter. A 4.88 m ASME/ANSI carbon steel pipe of 14-inch nominal bore,
schedule 80, with an external diameter of 355.60 mm, internal diameter of 317.50 mm and
thickness of 19.05 mm was selected for making up the shell of the unit. The heat exchanger
proportionality ratio (L / Ds) was calculated to be 13.72 .

(Sinnott, 2005) Stated that the proportionality ratio is mostly considered in heat exchangers
with removable tube bundles. These type of heat exchangers require additional vacant space in
their vicinity to allow the removal of the tube bundle for cleaning. Therefore, if the
proportionality ratio is considerably high, a considerable amount of industrial floor space is
needed. However, in fixed tube sheet heat exchangers, the tube bundle cannot be removed;
hence, the proportionality ratio is usually not considered as no industrial floor space is left
vacant to account for cleaning. For large proportionality ratios greater than 10, difficulty is
usually experienced in supporting the shell of the unit (Sinnott, 2005). The use of smaller tube
diameters less than the ¾ inch, which is recommended by (Kern, 1950) might have contributed
to the disproportionality of the unit.

Carbon steel was used to fabricate the shell of the heat exchanger. The corrosion inhibitors
present in Dowtherm fluid flowing on the shell side enables the fluid to be compatible (low

19
corrosion rate) with carbon steel. To prolong the lifespan of the shell, coating the inner and
outer surface using epoxy phenolic resins will be implemented as suggested by (Lancaster,
1983). Coating the shell on the outer surface serves to protect the unit from external corrosion
which may be caused by leakages from pipelines and other units in the vicinity.

Heat transfer coefficients (HTC3) in heat exchangers are functions of the flow regime across
these units. From the calculated Reynolds number of 15932.88 and heat transfer factor of 0.004
obtained from figure D 2 of appendix D, the tube side heat transfer coefficient of 2322.65 W/m2
K was obtained. Moreover, the shell side HTC is firmly dependant on both the flow regime
and baffle spacings. Optimising the baffle spacings to 0.42 of the outer shell diameter resulted
in a shell side HTC and shell side velocity of 2168.24 W/m2 K and 0.519 m/s, respectively.
The baffle spacing of 0.42 × Ds is within the range of (0.3 – 0.5) × Ds recommended by
(Sinnott, 2005). Moreover, the shell side velocity obtained was within the range of (0.3 – 1.0)
m/s recommended by (Sinnott, 2005). The distance between two adjacent baffles was 0.130 m,
and 25 % horizontal cut segmental baffles were used.

The values of the heat transfer coefficients presented here incorporated the viscosity correction
𝜇 0⋅14
factor (𝜇 ) of 0.98 and 1.09 for the tube and shell side, respectively. The wall temperature
𝑤

was determined to be 44.2 °C, and the viscosity of both the shell and tube side fluid were
obtained at this temperature. The slight deviation of the shell and tube side correction factors
from unity is inherent to the less viscous nature of both fluids as (Sinnott, 2005) stated that the
correction factor usually deviates from unity for highly viscous fluids. Therefore, Dowtherm
and the cumene-DIPB mixture can be regarded as low viscous fluids. The low viscous nature
of Dowtherm promotes its use in industrial heat exchangers. Table 3 below lists design
parameters for two units, one cooled using PAO and the other Dowtherm when all other
variables are kept similar.

Table 3: Comparison of design parameters for PAO and Dowtherm cooled STHE

Parameter PAO cooled STHE Dowtherm cooled STHE


HTC( Tube | Shell) (W/m2K) 1370.93 836.75 2322.65 2168.24
LMTD (°C) 26.76 32.28
Heat transfer area (m2) 49.30 28.51
Overall HTC (W/m2K) 221.90 331.11

3
In this text, “HTC” refers to the heat transfer coefficient

20
The life span of a heat exchanger may be significantly reduced by high-pressure drops. Using
the recommendation by (Sinnott, 2005) for pressure drops across heat exchangers, the tube side
pressure drop was restricted below 35 kPa as the viscosity of the mixture was 0.53 mPa-s at
52.5 °C. On the other hand, the shell side pressure drop was constrained in the range of 50 –
70 kPa as the viscosity of Dowtherm is 3.94 mPa-s at 20 °C. Therefore, the unit was optimised
to attain a total pressure drop of 31.67 kPa and 55.19 kPa on the tube and shell side,
respectively. The shell side pressure drop was estimated from Kern’s method. The shell of the
unit was made from a standard pipe with an inner diameter of 317.5 mm which is higher than
the numerically determined shell diameter of 309.25 mm. For a pragmatic design, the standard
pipe inner diameter of 317.5 mm was used in the calculation of the shell side pressure drop.
Predicting the pressure drop using the numerically determined shell diameter of 309.25 mm
will result in low-pressure drops during service as a larger shell diameter of 317.5 will be used
in the actual fabrication of the unit. Having a low shell side pressure drop during service may
result in a low shell side heat transfer coefficient. In addition, the pressure drop values
presented here are obtained from correlations that include the viscosity correction factor
𝜇 −0⋅14
(𝜇 ) of 1.02 in both the shell and tube side.
𝑤

The overall heat transfer coefficient was established from an iterative procedure where an
initial guess of 430 W/m2-K was used. (Sinnott, 2005) Suggested that the difference between
the assumed and calculated overall HTC should be less than 30 %. From the fouling resistances
of the shell and tube side fluid listed in table E 1 of appendix E, the overall HTC was
numerically determined to be 331.11 W/m2-K, which was 23 % less than the assumed value.

Using the Factorial method present by (Sinnott, 2005), the basic cost of the heat exchanger was
interpolated to be $ 18 000 using figure D 3 in appendix D. Accounting for inflation with a
cumulative price change of 35.92 % between 2004 and 2019 presented in (Alloth, L. C., 2013),
the basic cost of the unit was $ 24 465.6 in 2019. The pressure of the mixture from the column
was specified to be 1 bar; thus, the pressure factor was 1.0 as depicted in figure D 3 of appendix
D. Moreover, since the unit was fixed tube sheet, the type factor was 0.8. Incorporating the
pressure and type factor results in a total cost of $ 19 572.48. The currency exchange rate on this
day of 15 October 2019 is $ 1.0 = R 14.88; hence, the current cost of the heat exchanger is R
291 238.50 .

21
Conclusions

The final design of the heat exchanger met all the specifications stipulated by Process Design
Associates (Ltd) except for the proportionality ratio of the tube length to outside shell diameter
(L/Ds). Therefore, the use of tube diameters less than ¾ inch may result in a disproportional
heat exchanger. However, the incentive associated with using smaller tube diameters is the
small heat transfer area which is considered more valuable than the proportionality ratio from
the viewpoint of costs. Changing the cooling fluid from PAO to Dowtherm SR-1 on the shell
side of the unit increased the log mean temperature difference, and the heat transfer area was
reduced from 49.30 to 28.51 m2. Therefore, it was found that Dowtherm SR-1 fluid was a better
coolant than PAO fluid in shell and tube heat exchangers. Dowtherm SR-1 fluid was used as a
substitute for PAO because of its low viscosity and high thermal conductivity. Changing the
cooling fluid increased the shell side heat transfer coefficient from 836.75 to 2168.24 W/m2-
K. From this observation, the relationships postulated by (Mukherjee, 1998) for the heat
transfer coefficient as a function of thermal conductivity and dynamic viscosity were proven
to be valid. In this work, the baffle spacings, shell and tube side velocities and other design
parameters were set to be within ranges recommended by (Sinnott, 2005). Moreover, the heat
transfer area was also optimised to be within the range of 20 – 30 m2, which was a range
recommended by (Mukherjee, 1998) for heat exchangers using tube diameters less than ¾ inch.
It was then proven that using design parameters within ranges recommended in peer-reviewed
literature result in optimum designs. Therefore, these recommended ranges were considered to
be optimum ranges for use in heat exchanger design. The viscosity correction factor for
Dowtherm did not deviate significantly from unity; thus, Dowtherm was then classified as a
less viscous fluid which promotes its application in industrial heat exchangers. The
classification of Dowtherm as a less viscous fluid was made using the criterion established by
(Sinnott, 2005) regarding the deviation of the viscosity correction factor from unity. Fixed tube
heat exchangers have the lowest type factor 0.8 used for calculating the cost of heat exchangers.
Therefore, it was observed that fixed tube sheet shell and tube heat exchangers are the cheapest
to construct.

22
Recommendations

Fluid physical properties could be obtained from correlations developed as a function of


temperature to account for temperature variation in the heat exchanger. Accurate fouling
factors specific for the fluids used can be obtained experimentally for an accurate overall heat
transfer coefficient in the unit. Compare the results obtained with results from a professional
computer package such as Heat Transfer Research, Inc software (HRTI). This can be used to
check the suitability of Kern’s method used in this work. Lastly, Bell’s method could be used
to accurately estimate the shell side pressure drop.

Acknowledgements
I Would like to express my sincere gratitude to my lecturer, Mrs Anusha Singh, for her guidance,
support, and continuous encouragement throughout the design.

23
Appendices

Appendix A: Nomenclature
Table A 1: Nomenclature

Symbol Definition Units


Ao Heat transfer area m2
Atube The surface area of the tube m2
Ac Tube cross-sectional area m2
Cp,i Heat capacity of substance i J/kg.K
Db Tube bundle diameter m
de Equivalent shell diameter m
Ds Shell diameter m
di Tube inner diameter m
do Tube outer diameter m
hi Tube-side heat transfer coefficient W/m2.K
Rf,i Tube-side fluid Fouling factor W/m2.K
ho Shell-side heat transfer coefficient W/m2.K
Rof Shell-side fluid Fouling factor W/m2.K
jf Friction factor -
jh Heat transfer factor -
kf Thermal conductivity of fluid W/m.K
kw The thermal conductivity of the wall W/m.K
lb Baffle spacing m
K1 Bundle diameter coefficient -
𝑚̇𝑠 Mass flow rate in shell side kg/s
𝑚̇𝑡 Mass flow rate in tube side kg/s
n1 Bundle diameter coefficient -
Np Number of tube passes -

Nt Total number of tubes -

𝛥𝑃𝑠 Shell side pressure drop Pa


𝛥𝑃𝑡 Tube side pressure drop Pa
Q Heat duty W
Re Reynolds number -
o
Ti Temperature of substance i in tube side C
o
ti Temperature of substance i in shell side C
∆Tlm Log mean temperature difference o
C
𝑈𝑐𝑎𝑙 Calculated overall heat transfer coefficient W/m2.K
𝑈𝑎𝑠𝑠 Assumed overall heat transfer coefficient W/m2.K
𝑢𝑠 Shell-side fluid velocity m/s
𝑢𝑡 Tube-side fluid velocity m/s
𝜇i Dynamic viscosity of component i Pa.s
𝜌i Density of component i kg/m3
𝑣i Kinematic viscosity of component i m2/s

24
Appendix B
References

Alloth, L. C., 2013. U.S. Inflation Rates. [Online]


Available at: http://www.in2013dollars.com
[Accessed 15 October 2019].

DOW chemical company, 2010. DOWTHERM* SR-1 and DOWFROST*HD Heat Transfer
Fluids In Thermal Energy Storage. [Online]
Available at: http://www.hoodchemical.com/wp-content/uploads/2018/01/Engineering-
Operating-Guide-DOWTHERM-SR-1-4000.pdf
[Accessed 16 August 2019].

DOW chemical company, 2011. Engineering and Operating Guide. [Online]


Available at: http://www.hoodchemical.com/wp-content/uploads/2018/01/Engineering-
Operating-Guide-DOWTHERM-SR-1-4000.pdf
[Accessed 16 August 2019].

Engineering Toll box, 2003. Engineering tool box. [Online]


Available at: https://www.engineeringtoolbox.com/asme-steel-pipes-sizes-d_42.html
[Accessed 2 August 2019].

Gambill, W., 1959. "How to estimate mixtures viscosities". Chemical Engineering, Volume
66, pp. 151-152.

Ghajar, W. & Tang, A., 1994. Comparison of Hydraulic and Thermal Performance of PA0 and
Coolanol 25R Liquid Coolants. Washlngton, D.C., 6th AIANASME Joint Thermophysics and
Heat Transfer Conference.

Kern, D. Q., 1950. Process Heat Transfer. New York: McGraw-Hill.

Lancaster, J. .. F., 1983. Heat Exchanger Design Handbook. New York: Hemisphere
Publishing Corporation.

Lavine, A. S., Incropera, F. P., Bergman, T. L. & Dewitt, D., 2011. Fundementals of HEAT
and MASS Transfer. 7th ed. Hoboken: John Wiley & Sons. Inc.

Mukherjee, R., 1998. Effectively design shell-and-tube heat exchangers. Chemical


Engineering Progress, 94(2).

25
Serth, R. W., 2007. Process heat transfer principles and application. s.l.:Elsevier Science &
Technology Books.

Sinnott, R. K., 2005. Chemical engineering design. 4th ed. s.l.:Elsevier Butterworth-
Heinemann.

Appendix C: Piping and Instrumentation diagrams

Figure C 1: Section of the process flow diagram near heat exchanger E-607

Figure C 2: Heat exchanger mechanical drawing with and dimensions in mm

26
Figure C 3: Cross sectional view of heat exchanger tube layout with dimensions in millimeters

Appendix D: Figures and tables of sourced data

Figure D 1: Tube-Side friction factors (Sinnott, 2005)

Figure D 2: Tube-Side heat transfer factor (Sinnott, 2005)

27
Figure D 3: Shell and tube heat exchanger purchase cost for mid-2004

Figure D 4: Shell-Side heat transfer factors (Sinnott, 2005)

Figure D 5: Shell-Side friction factors (Sinnott, 2005)

28
Figure D 6: Bundle to Shell diameter clearance length (Sinnott, 2005)

Appendix E: Fluids physical properties and sample calculations

Temperature profiles for a counter current flow HE


90
Th,i
80
70
Temperature [°C]

𝑚̇ = 3.83 kg/s
60
50
40 Th,o Tc,o Hot
30 Cold

20
Tc,i 𝑚̇ = 4.60 kg/s
10
0
0,00 0,49 0,98 1,46 1,95 2,44 2,93 3,42 3,90 4,39 4,88
Tube length [m]

Figure E 1: Temperature profiles for shell and tube side fluid

29
Table E 1: Physical Properties of pure components retrieved using the NRTL model in Aspen® Plus Version 10.0 and (DOW chemical company, 2011)

Physical Property Cumene Diisopropyl benzene Cumene-DIPB mixture Dowtherm SR - 1


Mean Temperature (°C) 52.50 52.50 52.5 18.11
Specific heat capacity (J/kg K) 1880 1850 1878.21 3282
Density (kg/m3) 836 833 834.76 1074.5
Dynamic Viscosity (Pa.s) 5.3 x 10-4 5.6 x 10-4 5.30 x 10-4 0.00394
Thermal conductivity (W/m K) 0.117 0.112 0.117 0.367
Prandtl number 8.516 4.165 8.47 35.23
Fouling coefficient (m2 K/W) 0.0002 0.0002 0.0002 0.0002
Average Molecular Weight (g/mol) 120.19 162.20 120.24 40.04

Table E 2: Physical properties of fluids at wall temperature retrieved using the NRTL model in Aspen® Plus Version 10.0 and (DOW chemical company, 2011)

Physical Property Cumene Diisopropyl benzene Cumene-DIPB mixture Dowtherm SR - 1


Mean Temperature (°C) 44.20 44.20 44.20 44.20
Density (kg/m3) 843.3 839.0 838.46 1062.1
Dynamic Viscosity (Pa.s) 5.80 x 10-4 6.18 x10-4 6.16 x10-4 2.00 x 10-3
Kinematic viscosity (m2/s) 6.89 x10-7 7.37 x10-7 7.35 x10-7 1.883 x 10-6

Table E 3: Bundle diameter coefficients for triangular pitch (Sinnott, 2005)

Number of passes 1 2 4 6 8
K1 0.319 0.249 0.175 0.0743 0.0365
n1 2.142 2.207 2.285 2.499 2.675

30
Cumene-DIPB mixture property calculations

The specific heat capacity, density and thermal conductivity of the mixture were obtained from
group contribution of each component.

∑𝑁
𝑖=1 𝑥𝑖 𝜃𝑖

Where 𝑥𝑖 , 𝜃𝑖 are the mass composition and physical quantity of pure component 𝑖, respectively.

Therefore, 𝐶𝑃𝑚𝑖𝑥 = (0.998 × 1880) + (0.00135 × 1850) = 1878.74 J/kg - K

All other physical properties of the Cumene-DIPB mixture except the viscosity were calculated
as outlined above. The kinematic viscosity of the mixture was determined using the Gambill
method as outlined below (Gambill, 1959).

1⁄ 1⁄ 1⁄
3 3 3
𝜈𝑚𝑖𝑥𝑡𝑢𝑟𝑒 = 𝑥𝑎 𝜈𝑎 + 𝑥𝑏 𝜈𝑏 = 0.998 × (6.33 x 10-7)1/3 + 0.00135 × (6.72 x 10-7)1/3

𝑣𝑚𝑖𝑥𝑡𝑢𝑟𝑒 = 6.36 × 10-7 m2/s

Flowrates

From the specifications set by Process Design Associates (Ltd), the molar flow rate of the
distillate (Cumene-DIPB mixture) which was set to flow on the tube side of the heat exchanger
was 104.271 kmol/hr. Conversion of this flow rate into usable units is outlined.

𝑚̇𝑡 = molar flowrate x average molar mass = 104.271 × 120.24 = 12537.55 kg/hr

𝑚̇𝑡 = 12537.55 / 3600 = 3.48 kg/s

Incorporating an over-design factor of 10 % as explained in the discussion section :

𝑚̇𝑡 = 3.48 × 1.1 = 3.83 kg/s.

The flowrate of Dowtherm SR-1 fluid flowing on the shell side was not initially specified;
hence, it was set to be 20 % more than the flowrate on the tube side.

𝑚̇𝑠 = 𝑚̇𝑡 × 1.2 = 3.83 × 1.2 = 4.60 kg/s.

Heat duty

Neglecting heat losses to the surroundings, the hot fluid was used to determine the duty of the
heat exchanger.

𝑄ℎ𝑜𝑡 = 𝑚̇𝐶𝑝 𝑑𝑇

31
𝑄ℎ𝑜𝑡 = 3.83 ×1878.21 × (80 – 25) = 395726.11 W

From the negligible heat loss assumption, 𝑄𝑐𝑜𝑙𝑑 = 𝑄ℎ𝑜𝑡 = 395726.11 W.

Using the inlet temperature of 5 °C on the shell side, the outlet temperature was determined as
shown;

𝑄𝑐𝑜𝑙𝑑
Tout (shell) = + 𝑇𝑖𝑛
𝑚̇𝐶𝑝

Tout (shell) = 398.726.11/(4.60 × 3282) + 5 = 31.23 °C

Log mean temperature difference ( 𝛥𝑇𝑙𝑚 )

The log mean temperature difference for counter-current flow was also obtained.

(𝑇1 −𝑡2 )−(𝑇2 −𝑡1 ) (80−31.23)−(25−5)


𝛥𝑇𝑙𝑚 = 𝑇 −𝑡 = 80−31.23 = 32.28 °C
𝑙𝑛( 1 2 ) 𝑙𝑛(
25 − 5
)
𝑇2 −𝑡1

Since the heat exchanger is a 1-shell, 1-tube pass, the correction factor Ft = 1.

Total heat transfer area

From the assumed value of 430 W/m2 K for the overall heat transfer coefficient and the
calculated 𝛥𝑇𝑙𝑚 , the total heat transfer area was determined.

𝑄 395726.11
𝐴0 = = = 28.51 m2
𝑈𝑎𝑠𝑠 𝛥𝑇𝑙𝑚 430∗32.28

Tube layout and size

Carbon steel was selected as the material of construction of the tubes with a thermal
conductivity of 50 W/m K (Sinnott, 2005). 4.88 m length pipes were used with an external and
internal diameter of 21.034 mm and 6.4 mm, respectively. A triangular arrangement was used
with a pitch of (1.25 × Do = 26.29 mm).

Number of tubes

Heat transfer area of one tube (azimuthal surface area):

𝐴𝑡𝑢𝑏𝑒 = 𝜋𝑑0 𝐿 = 𝜋 × 0.021336 × 4.83 = 0.322 m2

ℎ𝑒𝑎𝑡 𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 𝑎𝑟𝑒𝑎 28.51


Number of tubes = = = 90 tubes
𝑎𝑟𝑒𝑎 𝑜𝑓 𝑜𝑛𝑒 𝑡𝑢𝑏𝑒 0.322

32
Tube side velocity

𝜋𝑑𝑖2 𝜋 × 0.00642
Cross-section area per pass (Ac) = × 𝑁𝑡 = × 90 = 0.00289 m2
4 4

𝑚̇ 3⋅83
Tube side velocity 𝑢𝑡 = = = 1.58 m/s
𝜌×𝐴𝑐 834⋅76 × 0.00289

Bundle diameter and shell diameter

Using k1 = 0.319, n1 = 2.142 and clearance of 0.012 m obtained from table E 3 and figure D 6.
1 1
𝑁 𝑛 90 2.142
Bundle diameter (𝐷𝑏 ) = 𝑑0 × (𝑘 𝑡 ) 1 = 0.021336 × (0.319) = 297.25 mm
1

Shell diameter (𝐷𝑠 ) = 𝐵𝑢𝑛𝑑𝑙𝑒 𝑑𝑖𝑎𝑚𝑡𝑒𝑟 + 𝑐𝑙𝑒𝑎𝑟𝑎𝑛𝑐𝑒 = 297.25 + 12.00 = 309.25 mm

Tube length / shell diameter ratio ( L / Ds) = 4.88 / 0.3556 = 13.72 (proportionality ratio)

From ASME document of standard pipes, a 14-inch nominal bore carbon steel pipe of 317.50
mm internal diameter and 355.60 mm external diameter pipe was used as the shell of the unit
(Engineering Toll box, 2003).

Tube side heat transfer coefficient

The heat transfer coefficient was first calculated without accounting for the viscosity correction
factor. From the value obtained, the tube wall temperature was determined, and the dynamic
viscosities of the fluids at wall temperature were then determined. The heat transfer coefficient
was then corrected by introducing the viscosity correction factor. This methodology was also
used in calculating the shell side heat transfer coefficient.

𝑢𝑡 𝑑𝑖 1.58∗0.0064
Reynold’s number (𝑅𝑒 ) = = = 15936.88
𝜈 6⋅35×10−7

L/Din = 4.83/0.0064 = 754.68

From figure D 2 the tube side heat transfer factor was obtained to be jh = 0.0040.

𝑁𝑢 = jh × 𝑅𝑒 × 𝑃𝑟 0⋅33 = 0.0040 × 15936.88 × 8.470.33 = 129.16

𝑘 0.117
ℎ𝑖 = (𝑑𝑖 ) × 𝑁𝑢 = × 129.16 = 2370.31 W/m2 K (without correction factor)
𝑖 0.0064

𝑈𝑎𝑠𝑠 (𝑇−𝑡)+ h𝑖 .𝑡 430 × (18.11 − 52.5) + 2322.65 × 18.11


𝑡𝑤 = = = 44.20 °C
ℎ𝑖 2322.65

From the calculated wall temperature, a viscosity correction factor of 0.980 was obtained.

33
ℎ𝑖 𝑑𝑖 𝜇 0⋅14
𝑁𝑢 = = 𝑗ℎ . 𝑅𝑒. 𝑃𝑟0.33 (𝜇 ) = 0.0040 ×15936.88× 8.470.33 × 0.98 = 126.57
𝑘𝑓 𝑤

𝑘 0.117
ℎ𝑖 = (𝑑𝑖 ) × 𝑁𝑢 = × 126.57 = 2322.65 W/m2 K (with correction factor)
𝑖 0.0064

Shell side heat transfer coefficient

Kern’s method was used to determine the shell side heat transfer coefficient.

Baffle spacing (𝑙𝑏 ) = 0.42 × 0.309 = 0.42 × 0.309 = 0.130 m

In calculating the area of cross-flow below, the actual shell diameter was used instead of the
calculated value.

(𝑝𝑡 −𝑑0 )𝐷𝑠 ×𝑙𝑏 (0.02667−0.021034) × 0.3556∗0.124


Area of cross-flow (𝐴𝑠 ) = = = 0.00825 m2
𝑃𝑡 0.02667

1.10
Equivalent diameter for triangular pitch (𝑑𝑒 ) = × (𝑃𝑡2 − 0.917𝑑02 )
𝑑0

1.10
𝑑𝑒 = × (0.026672 − 0.917 × 0.0210342 ) = 0.0151 m
0.021034

𝑚̇ 4.60
Mass velocity (𝐺𝑠 ) = = = 557.37 kg / m2 s
𝐴𝑠 0.00744

𝐺𝑠 557.37
Shell side velocity = = = 0.519 m/s (within specification)
𝜌 1074.5

𝑢𝑠 × 𝑑𝑒 0.519∗0.0151
Reynolds number (𝑅𝑒 ) = = = 2143.13
𝑣 3⋅67×10−6

25% cut horizontal segmental baffles were used, and a heat transfer factor (jh) of 0.0130 was
obtained from figure D 4.

ℎ0 𝑑𝑒 𝜇 0⋅14
𝑁𝑢 = = 𝑗ℎ × 𝑅𝑒 × 𝑃𝑟 0⋅33 × (𝜇 ) = 0.013× 2143.13×35.230⋅33 × 0.98 = 89.50
𝑘𝑓 𝑤

𝑘 0.367
(ℎ0 ) = (𝑑𝑓 ) × 𝑁𝑢 = × 104.95 = 2168.24 W/m2K (with correction factor).
𝑒 0.0149

Overall Heat transfer coefficient

𝑑
1 1 𝑑0 × ln( 0 ) 1 𝑑
𝑑𝑖
= ℎ + 𝑅𝑓, + + (ℎ + 𝑅𝑓, 𝑖 ) × ( 𝑑0 ) (E-19)
𝑈0 0 2 × 𝑘𝑤 𝑖 𝑖

0.021034 −1
1 0.021034∗ln( ) 1 0.021034
0.0064
𝑈𝑐𝑎𝑙 = {2585.01 + 0.0002 + + (1781.03 + 0.0002 ) × ( )}
2 × 50 0.0064

𝑈𝑐𝑎𝑙 = 331.10 W/m2K

34
𝑈𝑎𝑠𝑠 − 𝑈𝑐𝑎𝑙 430− 321.10
𝑈0 value % difference = × 100 = × 100 = 23.00 %
𝑈𝑎𝑠𝑠 430

Thus, the percentage difference between 𝑈0 values is within specification.

Tube side pressure Drop

Using a tube side friction factor of 0.0045 obtained from figure D 5, the pressure drop was
calculated as shown.

𝜇 −0⋅14
Viscosity correction factor (𝜇 ) = 1.02
𝑤

𝐿 𝜇 −0⋅14 𝜌𝑢𝑡
𝛥𝑃𝑡 = 𝑁𝑃 × (8 × 𝑗𝑓 × (𝑑 ) × (𝜇 ) + 2.5 ) ×
𝑖 𝑤 2

4.83 835.45∗1.58 32684.12


𝛥𝑃𝑡 = 1 × (8 × 0.0045 × (0.0064) × 1.021 + 2.5 ) × = = 31.67 kPa
2 1000

Shell side pressure drop

Using a shell side friction factor of 0.053 obtained from figure D 1, the shell side pressure drop
𝜇 −0⋅14
was obtained. (𝜇 ) = 1.02
𝑤

𝐷 𝐷 𝜌𝑢𝑠2 𝜇 −0⋅14
𝛥𝑃𝑠 = 8 × 𝑗𝑓 × ( 𝑑𝑠 ) × ( 𝑑𝑠 ) × × (𝜇 )
𝑒 𝑒 2 𝑤

0.305 4.88 1074.5∗ 0⋅5752


𝛥𝑃𝑠 = 8 × 0.053 × (0.0149 ) × ( 0.4 ) × × 1.021 = 55.19 kPa .
2

Therefore, both the tube and shell side pressure drops were within specification.

Costing

For the calculated heat transfer area of 28.51 m2, the basic cost was interpolated to be $18000
for carbon steel in the year 2004, using Figure D 3 in appendix D.
From the cumulative price change of 35.92 % between 2004 and 2019 (Alloth, L. C., 2013),
the price of the unit in this current year of 2019 is $18000 × 1.359 = $ 24465.56
The heat exchanger was a fixed tube sheet hence: Type factor = 0.8 and the cumene-DIPB
mixture was available at 1.01 bar which was in the range of 1 – 10 bar: Pressure factor = 1

𝑃𝑢𝑟𝑐ℎ𝑎𝑠𝑒 𝑐𝑜𝑠𝑡 = $ 24 465.56 × 0.8 × 1 = $ 19 572.48

The currency exchange rate on this day of 15 October 2019 is $ 1.0 = R 14.88

Therefore, the current cost of the heat exchanger is R 291 238.50 .

35

View publication stats

You might also like