EXERGETIC THERMOECONOMIC

HEAT EXCHANGER OPTIMIZATION MODEL

Liaquat Ali Khan

05-UET/PhD-ME-18

Department of Mechanical Engineering

Faculty of Mechanical & Aeronautical Engineering
University of Engineering &Technology
Taxila-Pakistan
July 2013
 EXERGETIC THERMOECONOMIC
HEAT EXCHANGER OPTIMIZATION MODEL

Liaquat Ali Khan

05-UET/PhD-ME-18

Department of Mechanical and Aeronautical Engineering

Faculty of Mechanical & Aeronautical Engineering

University of Engineering and Technology

Taxila-Pakistan
July 2013
ii
 EXERGETIC THERMOECONOMIC
HEAT EXCHANGER OPTIMIZATION MODEL

Author

Liaquat Ali Khan

05-UET/PhD-ME-18

A thesis submitted in partial fulfillment of the requirement for the

degree of

PhD Mechanical Engineering

Thesis Supervisor

Asst. Prof. Dr. Muhammad Shehryar

Thesis Supervisor‟s Signatures ________________________

_____________________________ ______________________________
External Examiner‟s Signatures External Examiner‟s Signatures
Prof. Dr. Bsharat Ullah Malik Prof. Dr. Abdul Ghafoor

Department of Mechanical Engineering

Faculty of Mechanical & Aeronautical Engineering

University of Engineering and Technology, Taxila, Pakistan

July 2013

iii
 Supervisor
Asst. Prof. Dr. Muhammad Shehryar

Members of Research Monitoring Committee

Prof. Dr. Arshad Hussain Qureshi

Prof. Dr. Zafar Ullah Koreshi

Prof. Dr. Shahab Khushnood

Foreign Research Evaluation Experts

Prof. Dr. Gary Marlin Sandquist, USA

Prof. Dr. Mikhael Gorokhovski, France

iv
 Foreign Expert Evaluation Report

Minor corrections (English grammar) made to abstract for dissemination. Contribution

was of engineering/economic significance for performing thermodynamic analysis and
optimization of heat exchangers using new methods of analysis.

Gary M. Sandquist

v
 Foreign Expert Evaluation Report

This thesis is focused on the complex problem of cost/design optimization in heat

exchangers. The major contribution of this thesis is two-fold. First, it is the new
application of evolutionary population-based algorithm in the context of optimization of
heat exchangers. Second, it is the economic analysis of improved designs, proposed in
the thesis.

The thesis is comprised of six chapters and six appendices. In the first chapter a
general description of heat exchangers, the basic formulation of zero-dimensional
thermodynamic models, and a general introduction to optimization methods are given. In
the second chapter, the evolutionary algorithm is introduced. On the basis of this
approach, the cost/design sensitivity analysis for shell and tube exchangers is performed
in the third chapter, for different liquids and for different operating parameters. A
considerable amount of results is assessed by comparison with existing approaches in the
literature. Chapter four is devoted to energy recovery optimization in the cross-flow
plate-fin heat exchanger. Here again in coupling with convective heat transfer relations
in the plate-fin exchanger, the Author applied and assessed the evolutionary algorithm in
order to minimize consumptions in the air-conditioning system. The similar strategy is
applied in five‟s chapter for the cross-flow tube-fin heat exchanger. Here also the
emphasis is put on the economic significance of design improvements proposed in the
thesis. The thesis is ended up with conclusion (chapter six)

In summary, the thesis provides a novel application of optimization methods to

heat exchangers. The proposed designs and their economic analysis are straightforward
to be applied in engineering.
I thus give a favorable opinion for the defense of the thesis of Mr. Liaqat Ali Khan.

Mikhael Gorokhovski

vi
 Dedicated

To my family, my sisters

For their devotion and prayers for my success

vii
 Acknowledgement

I am grateful for the opportunity to work with my supervisor , Dr. Muhammad Shehryar,

for the guidance and encouragement conducting this research and the experience gained

from the discussion on the topic. Also, his open, frank discussions concerning

professional approach were enlightening and appreciated.

I am also thankful to Dr. Ali Ghalban and Dr. Fathi Mahfouz for their able

guidance in the early and middle stages of my research. Without their guidance and help

this research will not be possible.

The commitment of the remaining committee members, Dr. Arshad Qureshi, Dr.

Zafar Ullah Koreshi and Dr. Shahab Khushnood also recognized. Their comments and

suggestion were instructive.

The Department of Mechanical and Aerospace Engineering is acknowledged for

the continued generous support. Special thanks to Dr. Muhammad Shahid Khalil for his

continuous guidance in administrative matter of my PhD degree program.

Finally, I would like to thank my family, who has endured through my education

and give love and encouragement. My sisters are appreciated for their endless

unassuming support, without which this point could have never been reached.

viii
 Research Work Publications

1. Liaquat A. Khan, Fathi Mahfouz, Thermoeconomic Lifecycle Cost

Optimization of an Annular Fin Heat Exchanger, Pakistan Journal of
Engineering and Applied Sciences, vol 11, pp 129-140, July 2012.

2. Liaquat A. Khan, Fathi Mahfouz, Thermoeconomic Lifecycle Energy

Recovery System Optimization For Central Air-Conditioning System
Using Evolutionary Technique, Mehran University Research Journal
of Engineering & Technology, Volume 31, No. 1, pp 13-28, January,
2012.

3. Liaquat Ali Khan, Dr. Ali El-Ghalban, Heat Exchanger Exergetic

Lifecycle Cost Optimization using Evolutionary Algorithms,
WSEAS Transaction on Heat and Mass Transfer , Issue 1, Volume 3,,
pp125-136. January 2008

4. Liaquat Ali Khan, Dr. Ali El-Ghalban, Heat Exchanger

Exergoeconomic Lifecycle Cost Optimization, 3rd IASME/WSEAS
Int. Conf. on Energy & Environment, University of Cambridge, UK,
Feburary 23-25, 2008.

ix
 Abstract

An investigation on the life cycle cost and annual cost of heat exchanger using
evolutionary optimization for both shell and tube and compact heat exchangers (plate-fin
heat exchanger and tube-fin heat exchanger) is done in this thesis. Cost includes capital
cost, pumping cost and exergy destruction cost. Pumping and exergy destruction costs are
the components of operating cost. Considering life cycle cost during the design phase of
thermal systems gives the design effort more worth. Optimized thermal parameters,
which give minimum cost, are important targets for both designers and users. The
optimization technique used in this study is evolutionary algorithm. Evolutionary
Optimization technique has recently experienced a remarkable growth [1]. A rapid
solution of the design problem enables to examine number of alternative solutions of
good quality, obtained using Evolutionary algorithms. Evolutionary algorithm gives the
designer more degree of freedom in the final choice compared to the traditional methods.
The code for thermal design of heat exchanger and for evolutionary optimization is
written in MATLAB®.

For shell and tube heat exchanger life cycle cost optimization procedure has been
developed. The objective function is to minimize the total cost, which includes capital
cost and the sum of discounted annual energy expenditures cost related to pumping and
exergy destruction. Effect of inflation on operating cost over the life of heat exchanger is
also considered. A reduction of 11.4% in capital cost and savings in operating cost is up
to 90.8% is obtained, with an overall decrease of life cycle cost up to 46.5%. Hence the
dominant factor in the life cycle cost of heat exchanger is the operating cost [2,3]. The
results obtained with Evolutionary Strategy (ES) is also compared with Genetic
Algorithm (GA), Direct Search (DS) and Simplex Method (SM) which are inbuilt
functions in Engineering Equation Solver (EES®) professional version. The results
obtained using ES are very close to GA and also comparable with DS and SM.
Sensitivity analysis with electricity prices is also performed. Parametric study of annual
cost with different baffle spacing ratios (BSR), Pitch Ratio (PR) and tube arrangement is

x
also carried out. The results are compared with literature. The results obtained in this
study are better as compared to the published data due to the use of better convective heat
transfer coefficient empirical relations and the optimization technique.

For gas to gas heat exchange plate-fin heat exchanger is designed and optimized
using evolutionary optimization. In this case the net energy recovered (energy recovered
- operating cost) for the life cycle of energy recovery system used with centrally air-
conditioning system is maximized. From the study it is observed that installation of
energy recovery system with a central air-conditioning saved significant amount of
energy and payback period is also less than a year. Sensitivity analysis with respect to
electricity energy prices is done to check the proper functioning of the algorithm.

Tube-fin heat exchanger is designed and optimized using evolutionary

optimization. These heat exchangers are good for the heat exchange between liquid and
gas. In the analysis a case study for the waste heat recovery system for the exhaust hot
gases is considered. In this case the life cycle cost of annular fin heat exchanger is
optimized by taking both pumping and exergy destruction cost. In this study simplified
relations for Colburn factor and friction factors are obtained. For the design of the heat
exchanger the thirteen standard geometries listed in [4] are used. By taking the range of
values for different parameters listed in [4] new annular fin geometry is found which
gives the minimum life cycle cost.

xi
 Table of Contents

Chapter 1 Introduction and Problem Statement ........................................................... 1

1.1 Shell and Tube Heat Exchanger ............................................................................... 1
1.2 Compact Heat Exchanger ........................................................................................ 1
1.3 Exergy and its concept ............................................................................................. 5
1.4 Exergetic Analysis ................................................................................................... 7
1.4.1 Exergy change in a heat exchanger ................................................................. 7
1.4.2 Control Volume Exergy Balance .................................................................... 8
1.5 Lifecycle Cost of Heat Exchanger ........................................................................... 9
1.6 Literature Review..................................................................................................... 9
Chapter-2 Evolutionary Computing .............................................................................. 15
2.1 Comparison of GA, ES and EP .............................................................................. 16
2.2 Evolutionary Strategy Algorithm ........................................................................... 18
2.2.1 Initialization .................................................................................................. 19
2.2.2 Recombination .............................................................................................. 19
2.2.3 Mutation ........................................................................................................ 21
2.2.4 Fitness Evaluation ......................................................................................... 22
2.2.5 Selection ........................................................................................................ 22
2.2.6 Termination Condition .................................................................................. 23
2.3 Pseudo-Code and Flow Chart of Program ............................................................. 24
2.4 Code Benchmarking............................................................................................... 25
2.5 Sensitivity Analysis ............................................................................................... 27
Chapter-3 Cost Optimization of Shell and Tube Heat Exchanger .............................. 29
3.1 Heat Exchanger Analysis ....................................................................................... 30
3.1.1 Sizing Problem .............................................................................................. 30
3.2 Rating Problem ...................................................................................................... 35
3.2.1 Bell-Delaware Method for Shell and Tube Heat Exchanger .......................... 35
3.3 Pressure Drop ......................................................................................................... 36
3.4 Pumping Power ...................................................................................................... 37
3.5 Objective Function Calculation ............................................................................. 37
3.6 Optimization Methodology .................................................................................... 38

xii
3.7 Optimization Problem Description ........................................................................ 39
3.7.1 Objective Function ........................................................................................ 40
3.8 Results and discussions .......................................................................................... 40
3.8.1 Case studies analysis ..................................................................................... 40
3.8.2 Case 1: Methanol – brackish water exchanger ........................................... 42
3.8.3 Case 2: Kerosene – crude oil exchanger ....................................................... 43
3.8.4 Case 3: Distilled water – raw water exchanger ............................................. 44
3.9 Sensitivity Analysis ............................................................................................... 45
3.10 Exergetic Analysis ................................................................................................ 47
3.11 Results and discussions based on exergetic analysis ............................................. 50
3.11.1 Case study 4: The optimization of shell and tube ................................... 50
3.11.2 Case study-5: A parametric study in the optimization of a shell and tube
heat exchanger .......................................................................................................... 51
3.12 Conclusion ............................................................................................................. 55
Chapter-4 Optimization of Air-Conditioning Energy Recovery System using Plate
Fin Heat Exchanger ....................................................................................... 57
4.1 Thermal Design of Plate-Fin Heat Exchanger ....................................................... 59
4.1.1 Sizing of a Single-Pass Cross Flow Exchanger ............................................ 59
4.1.2 Rating of a Single-Pass Cross Flow Exchanger ............................................ 66
4.2 Pressure drop .......................................................................................................... 70
4.3 Cost function .......................................................................................................... 73
4.3.1 Operating cost ............................................................................................... 73
4.4 Modeling Verification ................................................................................................ 77
4.5 Energy Recovery System ....................................................................................... 78
4.5.1 Types of Energy Recovery System ............................................................... 79
4.6 Optimization Methodology .................................................................................... 79
4.7 Results .................................................................................................................... 83
4.7.1 Case study analysis ..................................................................................... 83
4.7.2 Effect of different geometries on cost of energy recovered ....................... 84
4.7.3 Effect of fresh air inlet temperature on amount of energy recovered ........... 86
4.7.4 Effect of fresh air inlet relative humidity on amount of energy recovered ... 87
4.8 Optimization process ............................................................................................. 88
4.9 Payback period of energy recovery system ........................................................... 90
4.10 Capital recovery rate .............................................................................................. 91
4.11 Sensitivity analysis................................................................................................. 92
4.12 Conclusion ............................................................................................................. 93

xiii
Chapter-5 Lifecycle Cost Optimization of Waste Heat Recover System Using
Tube-Fin Heat Exchanger ............................................................................. 94
5.1 Tube-Fin Heat Exchanger ...................................................................................... 94
5.2 Tube-Fin Heat Exchanger Sizing Problem ............................................................ 95
5.2.1 Design Methodology..................................................................................... 97
5.3 Objective Function Calculation ........................................................................... 105
5.4 Optimization Methodology .................................................................................. 106
5.4.1 Optimization Problem Description ............................................................. 109
5.5 Results and Discussion ........................................................................................ 110
5.6 Conclusion ........................................................................................................... 115
Chapter-6 Conclusions and Recommendations for future work ............................ 116
Appendices ...................................................................................................................... 119
Appendix-A............................................................................................................. 120
Appendix-B ............................................................................................................. 141
Appendix-C ............................................................................................................. 142
Appendix-D............................................................................................................. 150
Appendix-E ............................................................................................................. 153
Appendix-F ............................................................................................................. 155
Appendix-G............................................................................................................. 160
REFERENCES............................................................................................................... 188

xiv
 List of Figures
Figure 1.1 A Shell and tube heat exchanger view ......................................................................................... 2
Figure 1.2 Compact heat exchanger view (Plate-Fin heat exchanger) [4] .................................................... 2
Figure 1.3 Counter-flow and Parallel flow arrangements and their temperature profiles ............................... 3
Figure 1.4 Cross flow arrangement with one fluid mixed and one unmixed [4] ............................................ 3
Figure 1.5 Cross flow arrangement with both fluids unmixed ....................................................................... 4
Figure 1.6 Interaction between the domains of energy, entropy and exergy .................................................. 6
Figure 1.7 Steady state process in a control region ........................................................................................ 8

Figure 2 1 The general scheme of an Evolutionary algorithm flow chart .....................................................16

Figure 2 2 Evolutionary Strategies flow chart ...............................................................................................17
Figure 2 3 Recombination techniques dominant recombination (at the bottom) and the intermediate
recombination (at the right side) [1] ............................................................................................20
Figure 2 4 Pseudo-code for Evolutionary Strategy ........................................................................................24
Figure 2 5 Flow chart showing the working of code using ES ......................................................................25
Figure 2 6 Comparison of minimum annual total cost using DS, GA, SM and ES .......................................26
Figure 2 7 Comparison of minimum annual total cost using GA and ES ......................................................27
Figure 2 8 A trend between number of generations and size of individuals ..................................................28

Figure 3. 1 Square and Triangular pitch tube layouts ....................................................................................33

Figure 3.2 Proposed EA optimization flow chart ..........................................................................................39
Figure 3. 3 Capital, Operating and Overall cost comparison for Case-1, Case-2 and Case-3 respectively
with [11] and [43] ........................................................................................................................45
Figure 3. 4 Capital, Operating and Overall cost comparison for Case-1, Case-2 and Case-3 respectively
with [2] ........................................................................................................................................45
Figure 3. 5 Program flow chart for sizing and rating problem and cost optimization ...................................49
Figure 3. 6 Comparison of AC-BSR for different tube arrangements ...........................................................53
Figure 3. 7 . Comparison of AC-BSR for different pitch ratio ......................................................................53
Figure 3. 8 Comparison of annual capital cost and exergy destruction cost ..................................................54

Figure 4.1 Basic construction of Plate fin Heat Exchanger Process Heat Transfer [44] ...............................57
Figure 4.2 A view of Cross-flow plate-fin Heat exchanger Process Heat Transfer [44] ...............................58
Figure 4.3 Plate fin surfaces examples: (a) Rectangular (plain) (b) rectangular (offset) ; (c) offset strip; (d)
triangular (plain) (e) triangular (perforated); (f) louvered; (g) wavy Process Heat Transfer
[44] ..............................................................................................................................................58
Figure 4.4 Surface basic characteristics for an offset strip fin surface (Kays & London [46]) .....................62

xv
Figure 4. 5 Proposed EA optimization flow chart .........................................................................................81
Figure 4.6 Program flow chart for sizing and rating problem and maximization to cost of energy recover .82
Figure 4. 7 Net saved energy value for different geometries of a 100 tons air-conditioning system ............84
Figure 4. 8 Net saved energy value for different geometries at different air-conditioning capacities ...........85
Figure 4. 9 Capital cost, Pumping cost, Exergy destruction cost, and net recovered energy for different
geometries at 100 tons of air-conditioning ..................................................................................86
Figure 4. 10 Net energy recovered value for different fresh air inlet temperatures of 100 tons of air-
conditioning .................................................................................................................................87
Figure 4.11 Comparison of net recovered energy for different fresh air relative humidity ...........................87
Figure 4.12 Net energy recovered per ton of air-conditioning for different air-conditioning capacities at
optimum conditions .....................................................................................................................89
Figure 4.13 Payback period of energy recovery system for different air-conditioning capacities.................90
Figure 4.14 Net amount of energy recovered along the life of energy recovery system ...............................90
Figure 4.15 Cost of conserved energy at different capacities ........................................................................91

Figure 5. 1A view of annular fin heat exchanger ..........................................................................................95

Figure 5. 2 Side view of heat exchanger to show tube arrangement ............................................................96
Figure 5. 3 Annular fin standard geometry [45] ............................................................................................99
Figure 5. 4 Proposed optimization flow chart .............................................................................................108
Figure 5. 5 Lifecycle cost of waste heat recovery radiator including exergy destruction cost ....................110
Figure 5. 6 Comparison of capital and operating cost (pumping + exergy destruction cost) ......................112
Figure 5. 7 Comparison of capital and operating cost (pumping cost) ........................................................113
Figure 5. 8 Optimized annular fin geometry ...............................................................................................114

xvi
 List of Tables

Table 3.1 Different parameters of fluids used in three cases .........................................................................41

Table 3. 2 A comparison of heat exchanger design .......................................................................................43
Table 3. 3 Sensitivity to the electricity price (reference exchange in case study # 1) ...................................46
Table 3. 4 Results for a different objective function (reference exchanger is case study # 1) .......................47
Table 3. 5 Data for the streams of STHE for case stude-4 ............................................................................51
Table 3. 6 Data for calculating computer program for case study-4 ..............................................................51
Table 3. 7 Design parameter determine by evolutionary algorithm compared with [3] ................................52
Table 3. 8 The design parameters comparison ..............................................................................................52
Table 3. 9 : Different component of cost comparison ...................................................................................52
Table 3. 10 Determined design parameters ...................................................................................................52
Table 3. 11 Determined design parameters ...................................................................................................53

Table 4.1 Values of n and m (source: Data from Shah [48]) .........................................................................71
Table 4.2 Entrance and exit pressure loss coefficients for (a) a multiple circular tube Core, (b) multiple-
tube at tube core, (c) multiple square tube core, and (d) multiple triangular tube core with
abrupt contraction and abrupt expansion [48] .............................................................................72
Table 4. 3 Operating conditions of PFHE (input data for the model) ............................................................77
Table 4. 4 The comparison of modeling output and the published results of Shah[48] and Ahmedi [50] ....78
Table 4. 5 Input data for the case study .........................................................................................................83
Table 4.6 Sensitivity to the electricity price ..................................................................................................92

Table 5. 1 Common parameters between present exchanger and 13 standard ...............................................97

Table 5. 2 Comparison of heat exchanger design considering pumping and exergy destruction costs with
true literature values ..................................................................................................................111
Table 5. 3 Comparison of heat exchanger design considering capital and pumping cost ............................112

xvii
 Nomenclature

A Surface area for plate-fin heat exchanger (m2)

a Surface area for shell and tube heat exchanger (m2)
a1 Numerical constant (€)
a2 Numerical constant (€/m2)
a3 Numerical constant
ACC Annual capital cost
B Baffle spacing in shell and tube heat exchanger (m)
BSR Baffle spacing ratio
b Plate spacing in plate-fin heat exchanger (m)
Cl Clearance in shell and tube heat exchanger (m)
Cp Specific heat at constant pressure (kJ/kg-K)
C Cost (€,$)
Cfuel Cost of fuel (€,$/kWh)
CP Pumping cost (€,$)
d Diameter (m)
dcc Center to center tube distance in depth direction (m)
F Temperature difference coefficient factor
f Friction coefficient
G Core mass velocity (kg/m2s)
h Convection heat transfer coefficient (W/m2K)
hcc Center to center tube distance in height direction (m)
Hr Annual operating time (h/year)
H Height of heat exchanger (m)
i Annual discount rate (%)
j Colburn factor
k Thermal conductivity (W/m K)
K1 Numerical constant

xviii
L Tube length (m)
L3 Length of plate-fin heat exchanger (m)
LCC Life cycle cost ($)
LHt Length of plate-fin heat exchanger on hot side (m)
LCt Length of plate-fin heat exchanger on cold side (m)
m Mass flow rate (kg/s)
mf Constant in fin efficiency
NER Net energy recovered ($)
n1 Numerical constant
Np Number of passes in shell and tube heat exchanger
NTU Net transfer units
Nu Nusselt number
ND Number of days (days)
NH Number of hours (hours)
OPC Total operating cost (€,$)
P Pressure (kPa)
Pr Prandtl number
PR Pitch Ratio
Pt Tube pitch (m)
Pf Fin Pitch (m)
Q Heat duty (kW)
Re Reynolds number
Relhum Relative humidity (%)
R Conductive fouling resistance (m2K/W)
S Cross-sectional flow area (m2)
T Temperature (K)
To Environmental temperature (K)
t Fin thickness (m)
TER Total energy recovered ($)
U Overall heat transfer coefficient (W/m2K)
v Velocity (m/s)

xix
V Volume (m3)
W Width of heat exchanger (m)
W Power (kW)

Greek

 Change
Δ Change
η Efficiency
μ Dynamic viscosity
α Heat transfer area/total volume (m-1)
ζ Cross –sectional area to frontal area ratio
 Heat transfer surface area density (m2/m3)
υ Kinematic viscosity
π Numerical constant
ρ Density (kg/m3)
ε Effectiveness
Ψ Exergy (kW)
λ Longitudinal conduction parameter
λ‟ Offspring individuals
μ‟ Parent individuals
ρ‟ Number of parents involved in the procreation of one offspring

Subscripts

capcost Capital cost (€,$)

C Cold fluid
C,in Cold fluid in
C,out Cold fluid out

D Days
xx
e exit
eq equivalent
Fuel Fuel
f Fresh air
fout Fresh air out
fr front
foul, shell Fouling shell side
foul, tube Fouling tube side
f,in Fresh air in
hin Hot fluid in
hout Hot fluid out
H Hot fluid
H,out Hot fluid out
H,in Hout fluid in
h Hydraulic
i Inner
ML Mean logarithmic
op Annual operating cost (€,$)
opD Discounted operating cost
o Outer
r Return air
rin Return air in
rout Return air out
r,in Return air in
s Shell
t Tube side
w Wall
y Year

xxi
Chapter 1 Introduction and Problem Statement

Heat exchangers are devices that transfer heat between two moving fluids. These
devices are used in power generation, heating, air conditioning, energy recovery systems,
motor vehicles, chemical and food processing. Heat exchanger is the most recognizable
heat transfer device and widely used in industry. Heat exchangers differ in both the
construction and flow arrangement. There are many types of heat exchanger designs;
each type has its own characteristics that make it suitable for a particular application. The
two broad categories of heat exchangers are
1. Shell and tube heat exchanger
2. Compact Heat Exchanger

1.1 Shell and Tube Heat Exchanger

The most common type of heat exchangers in industrial applications is the shell-
and-tube heat exchanger shown in Figure (1.1). Shell-and-tube heat exchangers contain a
large number of tubes (sometimes several hundred) packed in a shell with their axes
parallel to that of the shell. Heat transfer takes place as one fluid flows inside the tubes
while the other fluid flows outside the tubes through the shell. To maintain the uniform
spacing between the tubes and to force the fluid to flow across the shell to enhance the
heat transfer baffles are placed at equidistance inside the shell. Despite their widespread
use, shell-and-tube heat exchangers are not suitable for use in automotive and aircraft
applications because of their relatively large size and weight. Heat exchangers in which
all the tubes make one U-turn in the shell are called one-shell-pass and two-tube-passes
heat exchangers. Likewise, a heat exchanger that involves two passes in the shell and
four passes in the tubes in called a two-shell-passes and four-tube-passes heat exchanger.

1.2 Compact Heat Exchanger

Compact heat exchanger is designed to realize a large heat transfer surface area
per unit volume. The compact heat exchanger is shown in Figure (1.2). Heat transfer
area density β is the ratio of the heat transfer surface area to the volume of heat
exchanger. A heat exchanger with β>700m2/m3 is classified as being compact heat
exchanger. Examples of compact heat exchanger are car radiators (β =1000 m2/m3), glass
ceramic gas turbine heat exchangers (β=6000 m2/m3), the regenerator of a Sterling engine
(β=15,000 m2/m3), plate-fin heat exchanger (β=10,000-15,000 m2/m3) and the human
lung (β=20,000 m2/m3). Compact heat exchangers enable us to achieve high heat transfer
rates between two fluids in a small volume and, they are commonly used in applications
with strict limitations on the weight and volume of heat exchangers.

Figure 1.1 A Shell and tube heat exchanger view

Figure 1.2 Compact heat exchanger view (Plate-Fin heat exchanger) [4]

2
 Flow arrangements in a heat exchanger can be parallel, counter and cross flow. In
cross flow arrangement it may be one fluid mixed and one unmixed or with both fluids
unmixed. Figure (1.2) is showing the parallel and counter flow arrangement. Figures
(1.3a and 1.3b) are showing cross flow arrangement with one fluid mixed and one
unmixed and with both fluid unmixed respectively.

Figure 1.3 Counter-flow and Parallel flow arrangements and their temperature profiles [4]

Figure 1.4 Cross flow arrangement with one fluid mixed and one unmixed [4]

3
 Figure 1.5 Cross flow arrangement with both fluids unmixed [4]

The design of heat exchangers, in general, requires consideration of the heat

transfer occurring between the two fluids in addition to the mechanical energy needed to
overcome the frictional forces to move the fluids through the heat exchanger. These two
design criteria are generally the heat transfer and pressure drop. It is typically desired to
achieve large heat transfer yet maintain a small pressure drop. Large heat transfer rates
can be obtained by either having a large heat transfer area or having a large heat transfer
convection coefficient. Unfortunately, both of these conditions cause an increase in the
pressure drop. Since a large area gives more frictional loss resulting in an increase in the
pressure drop and the large flow rates to increase the convection coefficient would
likewise increase the pressure drop. There is a trade-off in these two design criteria; a
beneficial gain in one criterion is usually at expense of the other criterion and a
compromise must be established.
Usually heat exchanger thermal design is based on first law of thermodynamics.
An energy balance is done on the system for the heat transfer between the system and
environment. First law of thermodynamics is the law of conservation of energy,
according to which energy entering in thermal system in the form of fuel, flowing
streams of matter, electricity and from other sources is conserved and it cannot be
destroyed. Using first law of thermodynamics for energy balances does not provide
information about internal losses and the quality or grades of energy crossing the thermal
system boundary. On the other hand, second law of thermodynamics introduces the
concept of exergy in the analysis of thermal systems.
Exergy is a measure of the quality or grade of energy and it can be destroyed in
the thermal system. The second law states that part of the exergy entering a thermal
system with fuel, electricity, flowing streams of matter, and so on is destroyed within the

4
system due to irreversibilities. The second law of thermodynamics uses an exergy
balance for the analysis and the design of thermal system. The exergy balance is a
statement of the law of degradation of energy. Degradation of energy is due to the
irreversibilities of all real processes.
The design effort has more worth if life cycle cost of thermal system is considered
during the design phase. For the given constraints thermodynamic exergetic optimization
is a proven useful method for determining the optimal life cycle cost design of thermal
systems. Economic life cycle cost calculations based on first law of thermodynamics may
not be the better method for life cycle cost. By including second law of thermodynamics
(exergetic) analysis for the life cycle cost optimization will result in a thermodynamically
efficient and economically better heat exchanger design. Shell and tube heat exchanger,
plate fin heat exchanger (energy recovery system for air-conditioning system) and tube
fin heat exchanger (energy recovery for heating systems) life cycle cost and annual costs
are optimized using exergetic optimization.
There are two costs related to a heat exchanger, the capital (material and
manufacturing) cost and the operating cost (pumping and exergy destruction costs).
Optimization includes the minimization of both capital cost and running cost (minimize
their sum). The minimization of exergy destruction will reduce the operating cost of the
heat exchanger. The exergetic optimization is an attempt to reduce the energy
consumption and the energy lost to space, and hence the better control of environment.
The energy consumption in this sector represents considerable criteria for domestic,
commercial and industrial sector. Exergetic optimized heat exchanger will improve
energy consumption rate. This is beneficial for saving the energy reserves as well as the
most economical use of energy resources.

1.3 Exergy and its concept

First law of thermodynamics is an expression of the conservation of energy

principle. According to first law of thermodynamics energy is a thermodynamic property
and that during an interaction, energy can change from one form to another but the total
amount of energy remains constant. According to second law of thermodynamics energy
has quality as well as quantity, and actual processes occur in the direction of decreasing

5
quality of energy. High-temperature thermal energy is degraded because it is transferred
to a lower temperature body. To quantify the quality or “work potential” of energy in the
light of the second law of thermodynamics is explained using the properties entropy and
exergy.

Figure 1.6 Interaction between the domains of energy, entropy and exergy

The maximum useful work that can be extracted from a system until it reaches to
equilibrium with its environment is called exergy. Irreversibilities of a process caused
the exergy destruction. Second law analysis (or exergy analysis) is a powerful way of
optimizing complex thermodynamic systems. When all system and environment
processes are reversible then exergy remain conserved. For the irreversible system
exergy destruction occurred. Performing exergy analysis on a plant such as a chemical
processing plant or an entire power station, or a refrigeration plant, the thermodynamic
imperfections can be quantified as exergy destruction, which is wasted potential for the
production of work or wasted work
Information about the degradation of energy occurring in the process and to
quantify the usefulness or quality of the heat content in various streams leaving the
process as products, wastes, or coolant cannot be predicted using first law of
thermodynamics (FLT). The second law of thermodynamics (exergy) analysis
overcomes the limitations of the FLT. Exergy concept is based on first law of
thermodynamics as well as on second law of thermodynamics. Location of energy
degradation in a process can be indicated clearly using exergy analysis which may lead to
improved operation or technology. It quantifies the quality of heat in a rejected stream.

6
1.4 Exergetic Analysis

Cost sources at the component level can be obtained by combing the concept of
exergy with principles of engineering economy for the optimization of thermal systems
which is known as thermoeconomic analysis. Available energy destruction or exergy
loss represents the most important item for the exergetic analysis of thermal systems.
This loss in the available energy due to process irreversibilities are imposed by the
process boundary conditions and flow characteristics

1.4.1 Exergy change in a heat exchanger

For any fluid flowing through thermal system the exergy change is due to
temperature and pressure changes Van Wylen [5]. Equation-1.1and Euation-1.2 give the
exergy change for the hot fluid per unit time because of changes in temperature and
pressure respectively as it flows through the heat exchanger. Equation-1.3 represents the
total rate of exergy change for the hot fluid.
 T 
 H C PH (TH,out  TH,in )  To C PH ln  H,out
Δψ H,T  m  (1.1)
T 
  H,in 
 H,P  m
Δψ  H ν H ΔPH (1.2)

  T 
Δ ψ H  m H (ψ 2  ψ1 )H  m H CPH (TH,out  TH,in )  υHΔPH  ToCPHln  H,out 
T  (1.3)
  H,in 
Similarly, the exergy change for the cold fluid can be calculated using the change
in cold fluid temperature and pressure, as it flows through the heat exchanger.

  TC, out 
Δ ψ C  m C (ψ 2  ψ1 ) C  m C C PC (TC, out  TC, in )  υ C ΔPC  To C PC ln 
   
 (1.4)
  TC, in 

7
1.4.2 Control Volume Exergy Balance

Figure 1.7 Steady state process in a control region

At steady state, the exergy balance for the control volume within a thermal system
as shown in figure-3 per unit time takes the form;
0  ψ   ψ
 q, j   WCV
 i  ψ
e ψ
D (1.5)
j j e

The first term represents the exergy transfer due to the heat exchange, the second
term represent the work exchange, third and fourth terms represents the exergy exchange
with the incoming and outgoing fluid streams respectively and the last term is for exergy
destruction due to the process irreversibility. Explicitly equation-6 could be reformulated
as following;
.
T . .
0   (1  o ) Q j  Wcv   m
 i ψi   m
 eψe  ψD (1.6)
j Tj i e

In case of a heat exchanger

 0
Qj  W (1.7)
cv

The final equation for the exergy balance for a heat exchanger is
.
0  m
 i ψi   m
 e ψe  ψ
D (1.8)
i e

8
1.5 Lifecycle Cost of Heat Exchanger

Thermodynamically efficient design of thermal system might not be the design of

choice, because it may be too expensive to build and it is not cost effective to make a
profit. In the exergoeconomic optimization procedure it is desirable to maximize the
profit by minimizing the available energy loss per unit capacity (to minimize the per unit
cost).

1.6 Literature Review

Literature survey shows that the studies have done to perform the optimization of
component and thermal systems based on thermoeconomic (thermodynamics analysis
and economic analysis). It has been observed that life cycle cost is an important factor in
designing the thermal systems.
Yavuz Ozcelik [3] used the genetic algorithm (GA) to minimize the sum of
capital and exergetic cost of a shell and tube heat exchanger. The study also
implemented the parametric study using the genetic algorithm (GA).
Tozer et al. [6] optimized economically the HVAC system for an absorption-
cogeneration variable-air-volume system. They minimized the annual life cycle cost per
unit cooling capacity.
Summerer [7] estimated the cost of absorption system for its optimum design. He
maximized the coefficient of performance (COP) and minimized the cost. The amount of
heat exchanger surface area is fixed and varied the distribution of area. Finally he
selected the system with largest COP.
Tsatsaronis et al.[8] analyzed a power plant on the basis of exergoeconomic
analysis. They investigated the cost of power plant based on exergy destruction and
initial investment cost. They identify several changes to improve the cost effectiveness
of the power plant.
Zalewski et al. [9] designed the evaporative fluid cooler based on two
thermoeconomic objectives. The first objective was to minimize the total cost (capital +
running cost) by maximizing the heat transfer. The second objective was to find the
cooler design parameters to minimize the operating cost.

9
 Fowler and Bejan [10] estimated the optimum Reynolds number for external
flows over several simple geometries. Optimal parameters are found to have minimum
entropy generation. It is shown that the optimal Reynolds number scaled with the size of
the device, but with ever-increasing entropy generation. They found that the size of the
device should be increased to minimize the losses, but the tradeoff for larger size is the
cost. The tradeoff was exergy destruction or the cost of entropy generation by heat
transfer dissipation and the cost of size.
Antonio, Pacifico et al. [2] minimize the total cost of a heat exchanger (shell and
tube) by utilizing a genetic algorithm. Cost function included capital investment and the
sum of discounted annual energy expenditures related to pumping. They have compared
three different cases (methanol and see water, kerosene and crude oil, and distilled water
and raw water) and compared with the literature Kern [11] and Sinnot et.al. [12]. They
have shown that the genetic algorithm enables to examine a number of alternative
solutions of good quality and also gives rapid solution of the design problem. Number of
alternative solution gives the designer more degree of freedom in the final choice with
respect to traditional methods.
Swamee et al. [13] formulated a double pipe heat exchanger and optimized using
geometric programming with a single degree of difficulty. They optimized the cost of
double pipe heat exchanger by varying the inner diameter, outer diameter of heat
exchanger and the flow rate of the utility. The solution of the problem gives the optimum
values of the parameters for a given length of heat exchanger.
Selbas et al. [14] formulated a shell and tube heat exchanger for the estimation of
minimum heat transfer area required for a given heat duty which in turn makes the
overall cost of the heat exchanger (only the capital cost). They varied different design
parameters like, outer tube diameter, number of tube passes, tube layout, baffle spacing
etc. to the optimal design of shell and tube heat exchanger using genetic algorithm (GA).
Their objective function is based on total heat exchanger cost.
Ponce-Ortega et al. [15] used the Bell-Delaware method for the design of a shell
and tube heat exchanger and used the genetic algorithm (GA) to optimize the cost of heat
exchanger. Their optimization of cost dealt only with capital cost. They developed the

10
GA which gave the optimum solution even though the function is highly nonlinear
without convergence problem.
Philippe Wildi-Tremblay and Louis Gosselin [16] optimize the shell and tube heat
exchanger geometry by considering the pressure drops (shell and tube side) as constraints
by using GA. They also considered fouling in heat exchanger.
Arzu Sencan et al. [17] did the exergy analysis of lithium bromide/water
absorption systems. They analyzed the classics vapor absorption system using first law
and second law of thermodynamics. Their results showed that the exergetic efficiency of
the system decreased with increasing the heat source temperature whereas both cooling
and heating coefficient of performance (COP) of the system increased slightly with
increasing the heat source temperature.
Raj M. Manglik and Arthur E. Bergles [18] develop the relationships for heat
transfer and friction factor data for 18 offset strip fin surface. They show that j and f are
the function of geometry parameters. The range of validity is from laminar to turbulent.
Aytunc Erek et.al. [19] study the effect of fin geometry on heat transfer and
pressure dorop of a plate fin and tube heat exchanger. A computational fluid dynamics
(CFD) is done using Fluent software. They have studied effects of fin tube center
location, fin height, tube thickness, tube ellipticity, and distance between fins on heat
transfer between flue gas and water, and pressure drop.
Yonghan Kim et al. [20] did the experiment to have the optimal design of flat
plate finned-tube heat exchanger with large fin pitch. Their experimental data showed
that air side convective heat transfer increased as fin pitch reduced (more find per unit
length). Results also showed that there is a better heat transfer with staggered tubes.
George Tsatsaronis et al. [21] estimated the avoidable and unavoidable exergy
destruction and investment costs for a cogeneration system. The purpose of their
investigation was to know the avoidable part of exergy destruction and the avoidable
investment cost associated with a system component. They used the classical exergy
destruction method.
Cornelissen et al. [22] have considered for study water to water heat exchanger in
a city heating system. They analyzed the influence of the configuration of the heating
system, including the energy conversion on the optimization of the heat exchanger. They

11
combined the exergy analysis and life cycle analysis which lead to the lowest life cycle
irreversibility.
Hao Peng et al. [23] optimal design the plate-fin heat exchanger (PFHE) using
genetic algorithm combined with back propagation (BP) neural networks. The major
objectives were the minimum total annual cost for a given constrained conditions and the
minimum weight. They optimized the heat exchanger for minimum weight, smallest side
of the exchanger to have minimum capital cost, which the optimum pressure drops
accounting for the tradeoff between power consumption and heat exchanger weight.
Ponce-Ortega et al. [24] have designed and optimized the multipass heat
exchangers. For the design model they formulated the problem using the correction
factor for logarithmic mean temperature difference (FT). The optimization problem used
the minimization of the investment cost as the objective function. The optimization
algorithm used the number of shells as a convenient search variable to provide an
effective design method.
Cihat et al. [25] has carried out an analytical method to get optimum design of a
central heating radiator by varying geometrical and thermal parameters of the radiator.
Cost optimization based on capital and operating cost is missing in their optimization.
Dong-Keun Yang et.al. [26] proposed a method to improve the thermal
performance and to extend its operating time. They did it by optimizing the fin spacing
to optimize the heat transfer.
Sanjay K. Roy and Branko L. Avanic [27] they have developed a simple model
for heat transfer from a space radiator with latent heat storage. Two common
arrangements of tubes, i.e. square and equilateral triangular arrays were considered. It is
proved that more heat is transferred by using triangular arrangement.
Ganga et al. [28] optimize the surface area and pumping power required for a
tube-fin heat exchanger. They also study the effect of different materials used in radiator
like copper fins on copper tubes, brass and carbon steel tubes. The optimization
technique used was an analytical technique. Their optimization did not consider the
lifecycle cost of radiator.
Varun et al. [29] have given a new model for simulating air-to-refrigerant fin-and-
tube heat exchangers, with arbitrary fin sheet having varied geometric parameters. They

12
do not use the standard tube configuration. Instead their study is based on any arbitrary
configuration of tubes and fins. They optimized the surface area and pressure drop. Their
optimization was not based on cost and exergy destruction analysis.
Han-Taw et al. [30] have predicted the average heat transfer coefficient and fin
efficiency of annular-finned tube heat exchanger in natural convection using finite
difference method coupled with the least-square scheme and experimental temperature
data. They tried to optimize the heat transfer by maximizing the heat transfer coefficient.
Their optimization was also not based on lifecycle cost analysis.
Kundu et al. [31] have found out the optimum dimensions of fins for the tube-fin
heat exchangers for rectangular and equilateral triangular arrays of tubes. The
optimization is done by analytical method by taking first and second derivative by
maximizing heat transfer and minimizing the surface area. Their optimization did not
include the pumping cost and exergy destruction cost.
Arslanturk [32] found out the optimized dimensions of space radiators which
maximizes the heat transfer rate per unit radiator mass. His optimization problem did not
include cost optimization and the lifecycle cost of radiator.
Zhao-Gang et al. [33] used the Taguchi method for parametric study. They
selected fifteen samples from experimental database and optimized the design based on
flow depth, ratio of fin pitch and fin thickness and number of louvers. Their optimization
is based on minimum surface area and pumping power but it did not include the exergy
destruction cost.
From the above literature review it can be seen that previous techniques of
optimization involves the optimization of heat flux with minimum area [14][17] [22][13]
that is the capital cost and some take care of capital and pumping cost [7][8][10], and
some analyze the exergy destruction cost [15][3][20][21]. The analysis and optimization
done in this thesis differ from previous work in the formulation, solution of the problem
and area of applications. Although the previous techniques of designing and optimization
are present but these are not the direct approaches to develop the appropriate cost
function for a Shell and Tube heat exchanger, Plate-Fin heat exchanger and Tube-fin heat
exchanger for the lifecycle cost of heat exchanger. Shell and tube heat exchanger
lifecycle cost is optimized for pumping cost as well as for exergy destruction cost. Plate-

13
fin heat exchanger is optimized for energy recovery in air-conditioning application.
Similarly tube-fin heat exchanger is optimized for heat recovery system. Therefore, the
task is to develop appropriate cost function (capital cost + operating cost) for these types
of heat exchanger. Operating cost includes exergy destruction cost due to temperature
and pressure. Effect of inflation has been taken care of for the lifecycle cost.
Optimization is done using Evolutionary Strategy (ES) and results are compared
with the published data. In this study heat exchangers are optimized for minimum life
cycle cost considering exergy destruction (entropy generation) cost and capital cost with
present cost minimization. The heat exchanger is optimized for a given heat transfer rate.
Codes have been written in MATLAB® to design and optimize the lifecycle cost
of heat exchanger using evolutionary strategy (ES). Results obtained from ES are
validated with published data. Results are also validated with genetic algorithm (GA),
direct search (DS) method and simplex method which are inbuilt functions in EES®
professional software. Evolutionary strategy results are comparable with GA, DS and
Simplex method.
In this work, the method of solution is presented in chapter two which includes
the working of ES and comparison with GA. Shell and tube heat exchanger design,
analysis and optimization of its lifecycle cost function and results comparison is
addressed in chapter three, chapter four presents the solution methodology of plate-fin
heat exchanger for energy recovery system in centrally air-conditioning systems and
chapter five presents the design and optimization of tube-fin heat exchanger for heat
recovery system with results and discussion. A summary and the resulting conclusions
are given in chapter six in addition to recommendations for future work.

14
Chapter-2 Evolutionary Computing

Evolutionary optimization procedures derive inspiration from the Darwinian

Theory of evolution. It is an abstraction from the theory of biological evolution that is
used to create optimization procedures or methodologies, usually implemented on
computers that are used to solve problems. Evolutionary computation is based on natural
selection used in the field of study devoted to design, development and analysis in
problem solvers. This technique has been successfully applied to a wide range of
problems including tracking windshear, aircraft design, routing in communication
networks, robotics, air traffic, game playing, control design, scheduling making learning,
cost optimization etc.
There are three main independent implementation instances of evolutionary
algorithms (EAs), which are genetic algorithms (GAs), evolutionary strategies (ESs) and
evolutionary programming (EP). GAs, was developed by Holland [34] and thoroughly
reviewed by Goldberg [35]. Evolutionary strategies (ESs) were developed by
Rechenberg [36] and Schwefel [1] in Germany. Evolutionary programming originally
developed by L. J. Fogel et al. [37] and then refined by D.B. Fogel [38]. All these
algorithms have been proven capable of yielding approximately optimal solutions given
complex, multimodal, non-differential, and discontinuous search spaces.
There are different variations of evolutionary algorithms. All these variants have
common underlying idea which is same, that is given a population of individuals the
environmental pressure causes natural selection (survival of the fittest) and this causes a
rise in the fitness of the population. If a quality function is given to maximized of
minimized we create a set of candidate solutions, i.e., elements of the function‟s domain,
and apply the quality function as an abstract fitness measure. On the bases of fitness,
some of the better candidates are chosen to seed the next generation by applying
recombination and/or mutation to them. An operator called recombination applied to two
or more selected candidates (parents) and results one or more new candidates (children).
Mutation is applied to one candidate and results in one new candidate. By applying
recombination and mutation leads to a set of new candidates (offspring) that compete

15
based on their fitness function with the old ones for a place in the next generation. This
process can be repeated until a candidate with sufficient quality (a solution) is found or a
previously set computational limit is reached. In this complete process there are two
fundamental forces that form the bases of evolutionary systems.
i- Variation operators (recombination and mutation) create the necessary
diversity and thereby facilitate novelty, and
ii- Selection acts as a force pushing quality.
In consecutive populations the combined application of variation and selection
generally leads to improving fitness value. It is like a process as if the evolution is
optimizing, or at least “approximising”, by approaching optimal values closer and closer
over its course of execution. Hence a generalized evolutionary algorithm scheme is
shown in Fig.2.1.

Initialization Parent selection

Parents

Recombination
Population

Mutation

Termination
Offspring
Survivor selection
Figure 2 1 The general scheme of an Evolutionary algorithm flow chart

2.1 Comparison of GA, ES and EP

Although these algorithms are broadly similar, yet they have significant
differences. All methods operate on strings of fixed length, which contain real values in
EP and binary numbers in GA. All these algorithms incorporate a mutation operator: for
ESs and EPs mutation is the driving force. ESs and GAs also use recombination
operator, which is the primary operator for the GAs. All these three use a selection
operator which applies evolutionary pressure, either extinctive (in ESs and EP, the
operator determines which individuals will be excluded from the new population) or
preservative (in the GA the operator selects individuals for breeding). In genetic

16
algorithms and in evolutionary programming a probabilistic model is used for selection,
while evolutionary strategies use a deterministic selection.
In this thesis we are using evolutionary strategies for the cost optimization of heat
exchangers. Evolutionary strategy (ES) is usually faster and readily finds maximum or
minimum as compared to genetic algorithm (GA). In GA problem parameters can be
represented as bit strings which has computational problems. In ES the problem
parameters are real numbers and best suited for engineering problems [33-35]. The
objective of this thesis is to optimize the life cycle cost of heat exchanger (shell and tube,
plate-fin heat exchanger and tube-fin heat exchanger) which is an engineering problem
and has real numbers therefore; evolutionary strategy is suited for these kinds of
problems. The other benefit of using ES is that the solution obtained from this technique
takes less time as compared to GA. Fig-2.2 shows the flow chart of this method.

Initial Population

Recombination

Mutation

Fitness Evaluation

Selection

NO
Stop
Criterion?

YES

Figure 2 2 Evolutionary Strategies flow chart

17
 The process represented in Fig.2.2 is repeated until a stopping criterion is
achieved. This could be either a fixed number of generations or a minimum value of the
fitness function. Here we used fixed number of generations as stopping criteria because
convergence achieved within 100 generations.

2.2 Evolutionary Strategy Algorithm

The first step in Evolutionary Strategy is to generate a random population of

individuals each represents a potential solution. These individuals are also referred to as
the parent individuals, and their number is denoted by μ‟. In the recombination step, λ‟
pairs of these parent individuals are chosen and their values are recombined to produce λ‟
offspring individuals. The ρ‟ represents the number of parents involved in the
procreation of one offspring. The μ‟, λ‟ and ρ‟ are the strategy-specific parameters (the
mixing number) called exogenous strategy parameters, which are kept constant during the
evolution run. If ρ‟=1 then it is cloning and special ES cases without recombination. In
all other cases ρ >1 are strategies with recombination. In this thesis the code uses the
value of ρ is two. Recombination can be done either by exchanging randomly chosen
components of the parent pair or by taking average values. Once generated, Gaussian
noise is added to the solutions represented by these offspring individuals in the mutation
step. Gaussian noise is statistical noise which has its probability density function equal to
that of the normal distribution, which is also known as the Gaussian distribution. We
have assumed that the noise distribution in our ES is normally distributed. This step is
important to allow stochastic exploration of the design space. By adding normally
distributed noise the chance of selecting particular solution is evenly distributed. Every
solution has a fair amount of noise and no preference of one solution over the other.
Therefore the process is unbiased. The degree of noise added is a subject of interest in
Evolutionary Strategies Beyer and Schwefel [39]. The number of finite mutation steps or
generations is called reachability.
In this thesis the code written in MATLAB® uses the number of generations
which are taken to be 500. It was observed that the convergence is achieved within 100
generations but for the confirmation of the result it is preferred that no of generations

18
should be at least 500. Any of the selected individuals (parents) have to preference in ES,
and the variation operators should not introduce any bias Beyer and Schwefel [39].

2.2.1 Initialization

Evolutionary strategy is initiated by randomly generating a set of possible

solution. These individuals are also called parent individuals and their number is denoted
by µ. Every solution is referred to as an individual and the set itself is referred as
population. For reasonable good performance of optimization procedure it is necessary
that these individuals be scattered throughout the entire solution space. In all EA
applications the size of population is constant i.e. it is not changing during the
evolutionary search.

2.2.2 Recombination

If ρ > 2 (mixing parameter), the recombination is required. We are using ρ=2. In

GA the standard crossover uses two parents to produce two offspring. In standard ES
application, ES recombination operator to a parent family size ρ produces only one
offspring. The λ pairs of parent individuals are chosen and their values are combined to
produce λ offspring individuals. There are two standard methods of recombination used
in ES are
(i) Discrete recombination
(ii) Intermediate recombination
The discrete and intermediate recombination is shown in Fig.2.3. The upper left part of
the figure shows the mating process, which generate the parent family of size ρ. On the
right side of the figure intermediate ρ recombination is shown. At the bottom the
dominant (discrete) ρ recombination is represented.

19
Figure 2 3 Recombination techniques dominant recombination (at the bottom) and the intermediate
recombination (at the right side) [1]

 Discrete recombination

In Fig. 2.3 it is shown that vector a= (a1,….,ao) (object or strategy parameter

vector), the dominant ρ recombination produces a recombinant r= (r1,…,rD) by
coordinate-wise random selection from the ρ corresponding coordinate values of the
parent family.
(r)k = (amk)k with mk:= Random {1,…,ρ}
It means the component of the recombinant is determined exclusively by the
component of the randomly (uniformly) chosen parent individual mk that can be called as
“dominance”.
 Intermediate recombination
The intermediate recombination takes all ρ‟ parents equally into account. It
calculates the center of mass (centroid) of the ρ parent vectors am.

1
(r ) k :
'
 (a
m1
)
m k ….(2.1)

20
 This procedure is well defined for real-valued state space. Therefore, intermedian
recombination is used as the engineering problems are mostly in real-valued space. Our
problem is purely an engineering real value problem therefore we have used intermediate
recombination technique for recombination. For the application in discrete spaces
required an additional procedure as rounding or probabilistic rounding in order to map
back onto the discrete domain.

2.2.3 Mutation
The mutation operator is commonly a basic variation operator in ES. It is the
primary source of genetic variation. Up to now there is no established design
methodology, but some rules have been proposed Beyer [40] by analyzing successful ES
implementations and by theoretical considerations. Three parameters should be satisfied
in mutation process.
(i) Reachability
(ii) Unbiasedness
(iii) Scalability
 Reachability
The first requirement ensures that any other (finite) state can be reached within a
finite number of mutation steps or generations. This is also a necessary condition for
proving global convergence. In this thesis the objective function (optimization of life
cycle cost of heat exchanger) converges within hundred generations. For the
confirmation of the results we have taken the number of generations equal to 500.
 Unbiasedness
Variation (mutation) explores the search space. It does not use any fitness
information but the search space information from the parental population. Hence, any of
the selected individuals (parents) is no preference in ES and the variation operators
should not introduce any biasedness. The application of this principle leads to the normal
distribution (Gaussian distribution).
In this thesis because our problem is an engineering problem and lies in real valued
search space therefore normal distribution is taken to have an unbiased selection.
Gaussian noise is added in the selection step such that there will be no biasedness in the

21
selection process. An unbiased selection will produce a better equal opportunity to select
any one.
 Scalability
This requirement states that the mutation strength or the average length of a
mutation step should be tunable in order to adapt to the properties of the fitness
landscapes. The goal of adaptation is to ensure the evolvability of the ES system. It
means that the ES algorithm works in conjunction with the objective function. Since the
properties of the fitness landscape are determined by the objective function and the
variation operators, (i.e. the smoothness of the fitness landscape) are regarded as
fundamental requirement of efficient evolutionary optimization. The smoothness is
sometimes expressed in terms of causality Rechenberg [41]. Causality states that the
small changes on the genetic level should result on average in small changes in the fitness
values.

2.2.4 Fitness Evaluation

After the mutation step the fitness of each offspring individual is evaluated using
the given function f(x) and the µ best individuals are chosen as parent individuals for the
next generation. The value returned from the function is referred to as the fitness value
of the individuals. The main role of evaluation function is to represent the requirements
to adapt to mention function to be optimized. It is the basis for selection, and hence it
facilitates improvements. From the problem solving point of view, it represents the task
to solve in the evolutionary context. Evolution function is called fitness function in
evolutionary strategy. Mostly the problem solved by an EA is an optimization problem
which is sometimes called in the optimization context as objective function.

2.2.5 Selection
Evolutionary algorithm needs an operator which is goal oriented selection to
guide the search into promising regions of the object parameter space. Selection in ES is
almost alike animal breeding in which only those individuals with desired properties e.g.
high fitness value (objective function values) get a chance of reproduction [39]. Hence a
new population is obtained by deterministic process guaranteeing that only the µ best
individuals out of the selection pool of generation are transferred.

22
 There are two versions of selection technique, depending on whether or not the
parental population at (g) is included in the process or not.
(i) plus selection (µ+λ)
(ii) comma selection (µ,λ)
Comma section technique does not account for the old parents, i.e. selection
provides no search-relevant information and as a result the population performs a random
walk in the search.
 Plus Selection (µ+λ)
The plus selection also account for the old parents. The (µ+λ) shows that both the
parents and the offspring are copied into the selection pool which is therefore of size γ =
(µ+λ). There is no theoretical restriction on the offspring number λ. Cases like µ = λ or
µ > λ are also possible. Plus selection guarantees the survival of the best individuals
because it preserves the best individual such selection techniques are also elitist. Elitism
is sufficient condition a selection operator should obey in order to prove the ES‟s global
convergence property. Because plus strategies hold the elitism condition therefore,
parents can survive an infinitely long time span.
This technique is good to use in discrete finite size search spaces. Therefore, the
technique used in this thesis for selection process is (µ +λ) because the problem to
optimize (life cycle cost of heat exchanger) is a discrete finite problem and has finite
space. This technique has all the possible (parents and offspring) in the pool and the best
one either from the parent or from the offspring is taken. Hence the selection pool is vast
and the selection guaranteed the best to select.

2.2.6 Termination Condition

EA are stochastic and mostly there is no guarantee to reach an optimum, therefore

the condition to achieve the given precision (error ε) ε >0 might never get satisfied and
the algorithm may never stop. Commonly used options for this purpose are as follows.
(i) The maximum allowed CPU time elapsed.
(ii) The total number of fitness evaluation reaches a given limit.
(iii) For a given period of time (i.e, for the number of generations or fitness
evaluations).

23
 (iv) The population diversity drops under a given threshold.
The stopping criteria used in this thesis is the number of generations. The
program runs for 500 generation but it is observed that convergence is mostly achieved in
100 generations. Running the program for more no of generation is better to achieve
global minima.

2.3 Pseudo-Code and Flow Chart of Program

The Pseudo-code for the program made in MATLAB® scientific computing

environment is according to basic flow shown in Figure 2.2.

BEGIN
INITIALIZE population with random candidate solutions
EVALUATE each candidate
REPEAT UNTIL (number of generations) DO
1- SELECT parents
2- RECOMBINE pairs of parents
3- MUTATE the resulting offspring
4- EVALUATE new candidates
5- SELECT individuals for the next generation(µ+λ
END

Figure 2 4 Pseudo-code for Evolutionary Strategy

The flow chart of the computation done using evolutionary strategy (ES) is shown
in Figure 2.5. In the main program random numbers are generated, number of variable
parameters and its ranges are defined. In the next part of main program evolutionary
parameters like µ, λ, ρ, number of generations, type of recombination (intermedian) and
type of selection (µ + λ) are assigned. The main program calls the objective function that
is the heat exchanger thermal design and cost associated to capital and running cost is
calculated and comeback to the main program. The main program calls the evolutionary
strategy program which generates µ no of solutions and then it calls the recombination
program to produce λ offspring. This process is repeated for number of generations. The
control comes back to the main program which plots the objective function and the
variable parameters history with respect of number of generations. Final best cost
(minimum cost) and other parameters are displayed.

24
 Main Prog.
Define variation HX thermal
parameters, µ, λ, design and cost
ρ, no.of calculation
generations, type
of recombination,
type of selection

Call objective
function (HXG)

Display
obj
function
Variable
parameter
s and
plots
Main ES function

After mutation

Recombination
intermedian

Selection
(µ + λ)

Figure 2 5 Flow chart showing the working of code using ES

2.4 Code Benchmarking

Code is written in MATLAB® for the optimization of compact and shell and tube
heat exchanger cost (lifecycle or Annual cost). For this purpose ESs are used. For the
bench marking of the code the optimization is also done using Genetic Algorithm (GA),
Direct Search (DS) and Simplex method which are inbuilt optimization modules in
Engineering Equation Solver (EES®) professional version. The comparison shows that
the result obtained using the ES is comparable with the GA, DS and Simplex method.
ES, GA and DS methods, the objective function is first computed with a trial solution and
then the solution is sequentially improved base on the corresponding objective function
value till convergence.

25
 The problem is to minimize the annual cost (AC) of a shell and tube heat
exchanger. The cost function includes both the capital cost and the exergy destruction
cost.

AC  H  Cfuel (ψ ΔT  ψ ΔP )  FcrCC Hex (2.1)

And

i R (1  i R ) TL
Fcr  (2.2)
(1  i R ) TL  1
For the optimization, the range of continuous variable is given below.

10 ≤ m C ≤ 15

1.25 ≤ PR ≤ 1.5

0.2 ≤ BSR ≤ 1

The comparison of results for a shell and tube heat exchanger is shown in Fig. 2.6
and in Fig. 2.7. The expected outcome of the problem using ES, DS, GA and SM is
about $23280.
23500

Direct Search
23400
Genetic Algorithm
Simplex Method
23300 Evolutionary Algorithm

23200
ATC ($)

23100

23000

22900

22800

22700

22600
3 4 5 6 7
Tube Length (m)

Figure 2 6 Comparison of minimum annual total cost using DS, GA, SM and ES

26
 23500

23400
GA
EA
23300

23200
ATC ($)

23100

23000

22900

22800

22700

22600
3 4 5 6 7
Length of Tube (m)

Figure 2 7 Comparison of minimum annual total cost using GA and ES

It is clear from Fig. 2.6 and Fig. 2.7 that results obtained using ES and GA are
very close to each other. The average percentage difference between the GA and ES
value is less than 0.01%. It is because of that the ES and GA are both the two types of
EA. Both techniques are quite similar. The advantage of using ES is that it is simple,
uses real numbers and convergence is achieved in less time. ES also requires less
computational resources. ES are better for engineering problems because it is
deterministic in nature.

2.5 Sensitivity Analysis

In evolutionary strategies each generation is made up of certain number of

individuals and it runs for a number of generations which varies mostly from 100 to 500.
A sensitivity analysis is done between the number of individuals and the number of
generations for which convergence is achieved. It is observed that if sample size is large
then convergence is achieved in less number of generations and if sample size is small
then convergence is obtained in more number of generations. This trend follows up to
certain size of individuals. Hence, a particular number of individuals are required to
minimize the chance of local minima. On the other hand taking very large sample size
increases the number of generations required for convergence. A compromise is done
between the sample size and number of generations. The Fig.2.8 shows the trend
between number of generations and sample size.

27
 45

40

35
Number of Generations

30

25
Optimization result
of thesis case study
20

15

10

0
5 15 25 35 45 55 65 75 85
Number of Individuals

Figure 2 8 A trend between number of generations and size of individuals

In this thesis the number of individual are taken 35 and number of generations
are taken from 100 to 500. Number of individuals taken is 35 such that the program runs
for a suitable number of generations to reduce the chance of getting local minimum.
Taking very low no of individual convergence achieved in large number of generations
and no of individuals are not properly distributed. Similarly taking very large individual
size will make the number of generation quite low to obtain the convergence. The
convergence is achieved within 100 generations in most of cases discussed in this thesis.

28
Chapter-3 Cost Optimization of Shell and Tube Heat Exchanger

A shell and tube heat exchanger is a class of heat exchanger design that has a
wide range of industrial applications. These applications include oil refineries,
refrigeration, power generation systems, heating and air conditioning systems and many
chemical processes. As the name shows, shell and tube heat exchanger consists of a shell
(a large pressure vessel) with a bundle of tubes inside it. One fluid flows in the tubes,
and another fluid flows over the tubes (through the shell) transfer the heat between the
two fluids. Baffles are provided along the tube bundles that placed along its length to
direct the shell fluid to flow across the tubes. This cross flow motion induces more flow
turbulence and hence increases the heat transfer coefficient.
The advantages of shell and tube heat exchanger are
1) They are applicable for a wide range of operating pressure and temperatures
2) They have large ratio of heat transfer surface area to volume as compared to double-
pipe heat exchangers,
3) They are easy to manufacture in large variety of flow arrangements and sizes and
4) They can be easily assemble and disassemble.
The optimum design of shell and tube heat exchanger must aim to meet the above
advantages and minimize the life cycle cost of the exchanger.
The method used to get the optimum shell and tube heat exchanger design for the
minimum life cycle cost is to optimize first the capital cost and pumping cost. These
results are compared with three already published results. In second phase shell and tube
heat exchanger is optimized based on second law of thermodynamics (Exergetic analysis)
to get cost effective thermoeconomic model. Thermodynamically efficient heat
exchanger design based on first law of thermodynamics analysis may not be the best
method for economic life cycle cost calculations. By including second law of
thermodynamics (exergetic) analysis in the life cycle cost optimization will definitely
result in a thermodynamic efficient and also economically better heat exchanger design.

29
3.1 Heat Exchanger Analysis

In the first step, sizing analysis was done to get the design specification for a shell
and tube heat exchanger, when hot fluid inlet and outlet temperatures and their mass flow
rates are specified. The design specifications are calculated in such a manner that it
should fulfill the designed constraints. The design constraints are on tube outer diameter,
baffle spacing and shell diameter.
Once the heat exchanger is designed, it should be confirmed for its operating
condition, means it should fulfill temperatures and pressure drop specifications. If the
design specifications do not give the required temperatures then the specifications are
slightly varied to meet the required temperatures and pressures.
In heat exchanger analysis there are two types of problems:
1- Sizing Problem
2- Rating Problem

3.1.1 Sizing Problem

In the sizing problem inlet and outlet temperatures of fluids, mass flow rates and
other geometric specification are known. Hence, all the operating conditions are known
or can be calculated from the first law of thermodynamics. When inlet and outlet
temperatures are given and/or calculated, the log mean temperature difference (LMTD)
method can be employed to give the required design. The required design must fulfill the
following:
1) It must meet the process requirements (temperature and pressure drops) i.e., it
must effect the desired changes in the thermal condition of the process stream within the
allowable pressure drops, and it must continue to do this until the next schedule shutdown
of the plant for maintenance.
2) It must withstand the service conditions, which includes the mechanical stresses
of installation, startup, shutdown, normal operation, emergencies, maintenance and the
thermal stresses induced by the temperature differences.
3) It must be maintainable, which usually implies choosing a configuration that
permits cleaning-tube-side and/or shell-side, as may be indicated, and replacement of

30
tubes and any other components that may be especially vulnerable to corrosion, erosion
or vibration.
4) The heat exchanger should cost as minimum as possible such that it is consistent
with the above requirements.
The attempts to provide methods for calculating heat transfer coefficient and
pressure drop on shell-side were those in which correlations were developed based on
experimental data for a typical heat exchanger. Kern method is one of these methods,
which was an attempt to correlate data for standard exchangers by a simple equation
analogous to equations for flow in tubes. However, this method can only be applied for a
fixed baffle cut (25%) and do not adequately account for baffle-to-shell and tube-to-
baffle leakages.

3.1.1.1 Kern Method for Shell and Tube Heat Exchanger

Kern method is one of the well-known approaches for shell and tube heat exchanger
sizing process. In this approach, an appropriate size of a certain heat exchanger type is
selected to meet a certain specified hot and cold fluid inlet and outlet conditions. The
approach is based on log mean temperature difference (LMTD) method to compute the
required heat exchanger surface area, A.
Q
A (3.1)
UTLMTD F

Log mean temperature difference is defined as follows.

(ΔΔH  ΔTC )
ΔTLMTD  (3.2)
 ΔT 
ln  H 
 ΔTC 
Where, F is the temperature difference corrective factor. The value of factor F can be
calculated once we know RF and PF.
(THin  THout )
RF  (3.3)
(TCout  TCin )

(TCout  TCin )
PF  (3.4)
(THin  THout )

31
  1 P 
ln  
R 1
2
 1  PR 
F for NTP>1 (3.5)
R 1  2  P(R  1  R 2  1) 
ln  
 2  P(R  1  R 2  1) 

F=1 (for NTP=1)

U is the overall heat transfer coefficient and is computed as [2, 24].

1
U (3.6)
1 do  1 
 R sf   R tf  
hs di  ht 

Rsf and Rtf are the fouling factor on shell and tube sides respectively and are

depended on fluid type and operating temperature. hs and ht are the shell and tube sides

heat transfer coefficients respectively.

The tube side heat transfer coefficient ht is dependent on Re, Pr and di/L. For Ret
<2,300 Stephan and Preuβer [42] calculated ht as
 d 
 0.0677.(Re t .Prt . i )1.33 
Kt L
ht  3.657   (3.7)
di  d
1  0.1. Prt .(Re t . i ) 0.3 
 L 

While for 2,300 < Ret < 10,000 Gnielinski [43] used the following correlation

 ft 
 .(Re t 1000).Prt  0.67 

ht 
Kt  8 1   d i   (3.8)
di  ft   L  
1  12.7. (Prt  1)  
0.67

 8 

Where ft is the Darcy friction factor [43] given as

f t  (1.82log 10 Re t  1.64) 2 (3.9)

For the Ret more than 10,000 the following Sieder and Tate correlation [44] is
used.

32
 0.14
K 1  μ 
h t  t .0.027 Re t Prt 3  t 
0.8
(3.10)
do  μw 

The shell side heat transfer coefficient hs is calculated using the following
correlation of Kern [11] :
0.14
K 1  μ 
h s  t 0.36 . Re s . Prs 3 .  t 
0.55
(3.11)
de  μw 

In shell and tube heat exchanger, tubes are mostly arranged in a square or
triangular arrangement with different pitch as shown in Fig.(3.1). Hydraulic diameter
(de) is used for the flow diameter. [11, 44].

Figure 3. 1 Square and Triangular pitch tube layouts

=

The hydraulic diameter, de for square tube arrangement is calculated as:

4 . (Pt 2  πd o2 /4)
d eq  ; (3.12)
πd o
For triangular tube arrangement is calculated as:

4 . (0.43 .Pt 2  0.5. π . d o2 )

d eq  (3.13)
0.5 π d o

To calculate tube and shell side Reynolds numbers the flow velocities are
evaluated using the following relations
mt np
vt  . (3.14)
πd o
2
Nt
ρt
4

33
 ms
vs  (3.15)
a sρs

Where „as‟ is the shell side surface area and can be calculated as follows.

d s B . Cl
as  and Cl  Pt  d o (3.16)
Pt

Number of tubes can be computed using the following relation

n1
d 
N t  K1  s  (3.17)
 do 
Constants K1 and n1 are defined according to the number of passes and tubes
arrangement (square or triangular) as given in [44].
When the total heat exchanger surface area, A is known the necessary tube length
L is calculated as
A
L (3.18)
π do Nt

Heat transfer and the tube side fluid temperature (cold fluid temperature) are
calculated using Eq.(3.19) and Eq.(3.20).

Qm
 H C pH (THin  THout ) (3.19)
Q
TCout  (3.20)
 H C pH
m
After calculating outlet temperature effectiveness of heat exchanger is calculated using
Eq. (3.2).
(THin  THout )
ε (3.21)
(THin  TCin )
The above mentioned design procedure can be summarized as follows
1. Calculate number of tubes and shell hydraulic diameter using Eq (3.17)-and Eq
(3.12 & 3.13)
2. Calculate tube side and shell side velocities and their Reynolds numbers using Eq
(3.14)and Eq (3.15).

34
 3. Calculate tube and shell side heat transfer coefficients and the overall heat transfer
coefficient are calculated using Eq (3.7, 3.8, 3.9 & 3.11). and Eq (3.6)
4. Finally heat exchange length is calculated using Eq (3.18).
The whole procedure (steps 1-4) should be repeated for the new values of
optimization variables until minimum cost is obtained.
If heat exchanger length is known than number of tubes Nt can be calculated as follows.
A
Nt  (3.22)
π doL

3.2 Rating Problem

For rating problem length of tubes, number of tubes, shell and tube diameters and
other details are known. Once we know all the designed parameters the resulting outlet
temperatures can be determined. This is like using a heat exchanger given to us off the
shelf, where we know everything about it except for its operating conditions.

3.2.1 Bell-Delaware Method for Shell and Tube Heat Exchanger

Bell-Delaware method is complex but more accurate method as compared to

Kern‟s method which is simple [11]. In the Delaware method, the fluid flow in the shell
is divided into a number of individual streams. Each of these streams introduces a
correction factor to the heat transfer correlation for flow across a bank of tubes. Rating
analysis is performed using this method.
After calculating the shell side Reynolds number the ideal shell-side heat transfer
coefficient is calculated and finally the real shell-side heat transfer coefficient is found by
multiplying with different correction factors as given in Eq. (3.230.
h o  h id (J c J l J b J s J r ) (3.23)

Where,
Jc = Segmental baffle window correction factor
Jl = Baffle leakage effect correction factor
Jb = Bundle bypass effect correction factor
Js = Correction factor for variable baffle spacing at the inlet and/or outlet sections

35
 Jr = Correction factor for adverse temperature gradient in laminar flow

Overall heat transfer coefficient for clean surface is calculated as follows

1 1 1 d o ro ln(ro /ri )
   (3.24)
Uc ho hi di kw
The fouled heat transfer coefficient is calculated using the following relation
1 1 1 do r ln(r /r )
   R ft  o o i (3.25)
Uc h o hi di kw
Where, Rft is the fouling resistance on the tube-side.
Using the overall heat transfer coefficient, Log Mean Temperature Difference
(LMTD) calculation, the required surface area and eventually the length of the heat
exchanger can be calculated for the required inlet and outlet temperatures of hot and cold
fluids.

3.3 Pressure Drop

Next step is to calculate the pressure drop on tube and shell side. This gives the
amount of power required to overcome this pressure drop. This eventually contributes to
operating cost of a heat exchanger. Tube side pressure drop is computed as the sum of
distributed pressure drop along the tubes length, pressure losses in elbows and in the inlet
and outlet nozzles. Following relations give the pressure drop in tubes and in tube bends.

ρ t v 2t L f t np
ΔPt  (3.26)
2d i
ρ t v 2t p f np
ΔPb  (3.27)
2
Where, pf is constant having different values in the literature. Sinnott [45]
assumes pf = 2.5, while Kern [11] assumes pf = 4. In case studies, pf = 2.5 is used.

Pressure drop on shell side is calculated using the following equation.

36
 ρ s v s2 L d s
ΔPs  fs (3.28)
2 B de
Where, f s  exp(0.576  0.19 ln(Re s )) is the friction factor.

3.4 Pumping Power

Pressure drops on tube and shell side should be compensated and hence it is
required to calculate tube side and shell side pumping power required overcome the
pressure drops.

 W
W  W

P t s

  1  m t (P ΔP )  m s ΔP 
W
η  ρ t
t b s (3.29)
ρs 

3.5 Objective Function Calculation

In first phase of this chapter heat exchanger life cycle cost includes, capital cost
and operating cost. Capital cost is the sum of material and manufacturing or construction
cost, while the operating cost consists of pumping cost only. Once the total operating
cost of the heat exchanger for entire life is formulated the present cost estimation due to
inflation rate is also included.

C TOT  C capcost  C opD (3.30)

The capital investment Ccapcost is computed as a function of the exchanger surface

area, adopting Hall‟s correlation [44].

Ccapcost  a1  a 2Aa3 (3.31)

Where a1=8000, a2=259.2 and a3=0.91 for exchanger made of stainless steel for
both shell and tubes.

37
 The total discounted operating cost related to pumping power required to
overcome friction losses is computed using the following equation

ny C op
C opD  
k 1 (1  i) TL (3.32)

Cop  W . Cfuel ..Hr (3.33)

Where TL, Cfuel and Hr are the total life of heat exchanger in years, the cost of fuel and
numbers of hours of operation per year..

3.6 Optimization Methodology

Design of optimum heat exchanger includes the following steps:

 based on required amount of heat transfer, the surface area and other design
specifications is estimated, assuming some initial values.
 capital cost, operating cost and the total cost (objective function) is calculated
 After this, by using the optimization algorithm the new set of values for the
design parameters are selected.
 the previous steps should be repeated until a minimum of the objective function is
obtained.
In addition to the optimizing algorithm, heat exchanger lifecycle optimizing model is
modified to include a comprehensive thermal and hydraulic design procedure. Thermal
design of heat exchangers is to calculate an adequate surface area to handle the thermal
duty for the given operating conditions whereas, the hydraulic analysis determines the
pressure drop of fluids flowing in the system, and consequently the necessary pumping
power to maintain the flow is calculated.
Design parameters are the mass flow rates for hot and cold fluids and their
temperatures at inlet and outlet (mt, ms, Tti, Tto, Tsi, Tso). Other parameters are found
using energy balance. Fixed parameters are assigned by the user like, tube arrangement
pattern (triangular or square), number of tubeside passes (1,2,3,4,…), thermophysical
properties of both fluids and the fouling resistances (Rsf, Rtf). The optimization

38
parameters, whose values are iterated, are tube outside diameter (do), shell diameter (ds)
and baffle spacing (B).
Based on actual values and the design specifications, fixed parameters and current
values of optimization parameters, values of heat transfer coefficients for shell and tube
side (hs and ht) and the overall heat transfer coefficient, U are calculated. After this, the
surface area (A), number of tubes (Nt), length of heat exchanger (L) and shell and tube
side velocities (vs, vt) are calculated. By calculating pressure drops on both shell and
tube side pumping cost is calculated. Surface area (A) is used to calculate the capital
cost. Pressure drop gives the power requirement. Based on power requirement pumping
cost is calculated. The inflation effect is also accounted for the life of heat exchanger.
This process is iterated until a minimum of the objective function is achieved within the
given convergence criteria. The whole process and sequence of iterations is shown in
Fig. 3.2.
Specifications
(Fixed parameters etc)

Updated optimization
variables Fixed parameters
HX design routine
(Tube pattern, NP,etc)

Design parameters
(meeting specifications)

Evaluation of
capital cost

Evaluation of
exergy cost

EA algorithm

Figure 3.2 Proposed EA optimization flow chart

3.7 Optimization Problem Description

The purpose of optimization procedure is to find out the minimum value of life cycle
cost, which represents the objective function, subjected to the given constrains.

39
 a. Specified values for inlet and outlet temperatures of hot fluid, (THin and THout)
b. Specified value for hot fluid mass flow rate, ( m t in kg/s)
c. Specified value of inlet temperature of cold fluid, (TCin)
d. Tubes layout (Triangular or Square)

Optimization process is carried out through an evolutionary technique and the

results are verified by the published results Caputo [2]. The method used for the
optimization is evolutionary strategy described in section 2.2. In the evolutionary
technique each generation is made of 35 individuals. The maximum number of
generation is set to 500. However, in the test, convergence was always obtained within
100 generations.

3.7.1 Objective Function

The objective of the optimization procedure is to determine the values of tube

outside diameter baffle spacing and shell diameter (do, B and ds) such that for given
operating conditions, heat exchanger life cycle cost is minimized. The fitness function is
therefore, defined as

Ctot= Life cycle cost

Subject to,
d s  [d s min d s max]
do  [do min do max]
B  [ B min B max]

3.8 Results and discussions

Results include two types of results one for cost without exergetic analysis and
the other with exergetic analysis. A computer simulation program developed in
®
MATLAB is used to optimize the cost using evolutionary strategy.

3.8.1 Case studies analysis

To check the effectiveness of the said approach some relevant case studies taken
from the literature, in order to have reliable reference of sizing data for the sake of

40
comparison. In these cases di=0.8do. Operating data used for a three case studies is
shown in Table 3.1.
Case #1: distilled water – raw water exchanger, duty 0.46 (MW)

Case # 2: kerosene – crude oil exchanger, duty 1.44 (MW)

Case # 3: methanol – brackish water exchanger, duty 4.34 (MW)

The design specifications for the above three cases presented in Table-3.1 were
supplied to the evolutionary optimization algorithm and the optimized parameters are
compared with the original designed solution, which is given in Table 3.2. The cost of
both original and the present EA solutions are computed and presented in Table 3.2.
The upper and lower bounds were imposed for the optimization variables, shell
diameter, baffle spacing and tube diameter (Ds, B, do).
0.1m ≤ Ds ≤ 1.5 m

0.05m ≤ B ≤ 0.5 m

0.015 m ≤ do ≤ 0.5 m

Table 3.1 Different parameters of fluids used in three cases

Mass T T Cp k Rfouling
ρ μ
flow input output (kJ/kg (W/m (m2
(kg/m3) (Pa s)
(kg/s) (ºC) (ºC) K) K) K/W)
Case 1
Shellside:
27.80 95.0 40.0 750 2.84 0.00034 0.19 0.00033
methanol
Tubeside:
68.90 25.0 40.0 995 4.20 0.00080 0.59 0.00020
sea water
Case 2
Shellside:
5.52 199.0 93.3 850 2.47 0.00040 0.13 0.00061
kerosene
Tubeside:
18.80 37.8 76.7 995 2.05 0.00358 0.13 0.00061
crude oil
Case 3
Shellside:
22.07 33.9 29.4 995 4.18 0.00080 0.62 0.00017
methanol
Tubeside:
35.31 23.9 26.7 995 4.18 0.00092 0.62 0.00017
sea water

41
 Discounted operating cost was computed with TL = 10 years, annual discount rate
i= 10%, energy cost Cfuel = 0.12 €/kWh and an annual amount of work hours H = 7000
hour/year.

3.8.2 Case 1: Methanol – brackish water exchanger

This case study was taken from Kern [11] and Caputo [2]. Two tube passes with
triangular pitch pattern and one shell side pass characterize the exchanger structure. The
same structure is retained in EA.
In this case, a significant increase of about 55.5% of the heat exchanger surface
area was observed with corresponding increase of capital investment cost (+ 25.5%) as
compared to [44]. An increase of about 16% in heat exchanger surface area was
observed with an increase of 8.4% in capital cost as compared to Caputo [2]. This was
due to significant increase in shell diameter and the number of tubes, even though there is
significant reduction in tube length and minor decrease in tube diameter. There is a
significant reduction in tube and shell side velocities and pressure drops, which
drastically reduce annual operating cost for about 90.8% as compared to Kern [11] and
slightly higher as compared to Caputo [2].

This saving in operating cost drastically reduces the effect of higher capital
investment allowing a marked reduction of about 46.5% in total costs compared to Kern
[43] and slightly higher cost as compared to Caputo [2]. If overall cost is considered
there is a differential increase in capital investment and significant decrease in operating
cost. We get an increase of 4226 € and a decrease of annual operating cost of 4057 €.
Hence, the payback period of this increased capital investment is 1.04 years (12.5
months). Capital cost is 1612 € and 137 € higher if compared with Caputo [2].

Appendix-C is showing the convergence. It is showing the convergence of the

best fitness values with respect to the generations. It is shown clearly from the Figure C-
1 that as the generations pass through different generations, the best fitness values of the
individuals are minimized. Figures C-2, C-3 and C-4 show the optimal values of do, B
and ds. In figures each cross indicates the individuals.

42
Table 3. 2 A comparison of heat exchanger design
Case study # 1 Case study # 2 Case study # 3
Literature Literature Present Literature Literature Present Literature Literature Present
values[43] values[2] values[11] values[2] values[11] values[2]
Ds(m) 0.894 0.830 0.692 0.539 0.630 0.649 0.387 0.620 0.480
B (m) 0.356 0.50 0.50 0.127 0.120 0.114 0.305 0.440 0.50
do (m) 0.020 0.016 0.015 0.025 0.020 0.015 0.019 0.016 0.015
L (m) 4.830 3.379 4.313 4.880 2.153 1.487 4.880 1.548 2.95
Pt (m) 0.025 0.020 0.019 0.031 0.025 0.019 0.023 0.020 0.020
Cl (m) 0.005 0.004 0.00375 0.006 0.005 0.00375 0.004 0.004 0.00391
Nt 918 1567 1170 158 391 796 160 803 500
Vt (m/s) 0.75 0.69 0.67 1.44 0.87 0.537 1.76 0.68 0.735
Ret 14,925 10,936 10,000 8,227 4,068 17,916 36,400 9.487 10,000
Prt 5.7 5.7 5.695 55.2 55.2 5.645 6.2 6.2 6.203
ht 3812 3762 9661 619 1168 3451 6558 6043 9949
(W/m2K)
ft 0.028 0.031 0.031 0.033 0.041 0.027 0.023 0.031 0.031
ΔPt+ ΔPb 6.251 4.298 6.17 49.245 14.009 3.35 62.812 3.673 5.35
(kPa)
as (m2) 0.0320 0.0831 0.06916 0.0137 0.0148 0.0148 0.0236 0.0541 0.0491
De (m) 0.014 0.011 0.011 0.025 0.019 0.015 0.013 0.015 0.0110
vs (m/s) 0.58 0.440 0.536 0.47 0.430 0.438 0.94 0.41 0.452
Res 18,381 11,075 12,825 25,281 18,327 13,807 16,200 8039 6362
Prs 5.1 5.1 5.11 7.5 7.5 7.6 5.4 5.4 5.39
hs 1573 1740 1970 920 1034 1173 5735 3467 4274
(W/m2K)
fs 0.330 0.357 0.295 0.315 0.331 0.291 0.337 0.374 0.337
ΔPs 35.789 13.267 18.01 24.909 15.717 13.86 67.684 4.365 9.029
(kPa)
U 615 660 775.72 317 376 380.7 1471 1121 1276
(W/m2K)
Sa (m2) 278.6 262.8 237.72 61.5 52.9 55.77 46.6 62.5 72.46
Ccapcost 51,507 49,259 45,658 19,007 17,599 18,067 16,549 19,163 20,775
(€)
Cop 2111 947 1149 1304 440 161 4466 272 409
(€/year)
CopD (€) 12973 5818 7766 8012 2704 1088 27,440 1671 2765
CTOT (€) 64,480 55,077 53,425 27,020 20,303 19,154 43,989 20,834 23,539

3.8.3 Case 2: Kerosene – crude oil exchanger

This case study was taken from Kern [11] and Caputo [2]. This is a heat
exchanger of four tube passes with square pitch pattern and one shell side pass. The
same configuration in maintained in EA approach.
In this case, a small decrease of about 5.7% in heat exchanger surface area was
observed as compared to Kern [11] and slightly higher than Caputo [2]. This was due to
increase of shell diameter and a strong increase in number of tubes but a significant

43
decrease in both tube diameter and length. Therefore, capital investment cost was
reduced by 5% as compared to [11] and comparable with Caputo [2]. There is a marked
decrease in flow velocities and pressure drops, which reduces the annual operating cost
by 87.6% as compared to [11] and about 63.4% as compared to [2]. This leads to a net
reduction in total cost for about 29.1% as compared to [11] and 5.66% as compared to
Caputo [2].

3.8.4 Case 3: Distilled water – raw water exchanger

This case was taken from Sinnott [43] and Caputo [2]. The design of shall and
tube heat exchanger in this case has two tube side passages (triangular pitch pattern) and
one shell side passage. The same architecture was retained in EA approach.
In this case, a slight reduction in heat exchanger surface area resulted due to
reduction in shell diameter, tube diameter and length of heat exchanger even though there
is an increase in number of tubes. Hence, capital investment was decreased
correspondingly by 11.4% as compared to Sinnott [43] and 7.3% as compared to Caputo
[2]. Shorter shell and higher number of tubes reduce the flow velocities, which leads to
decrease in pressure drop. Therefore, annual operating cost decreased significantly for
about 45.6% compared to [44] and 21% higher as compared to [11]. The net effect of
reduction in capital investment and operating cost led to a total cost reduction by 17.1%
compared to [43] and 3% as compared to [11].
An overall cost comparison (capital investment and operating cost) for the three
case studies is shown in Fig. 3.3.

44
 70000

60000
12973

Ccapcost
CopD
50000 7766

40000
Cost (Euros)

30000 27440

51507
45658 8012 2765
20000 1088

10000 20775
19007 18067 16549

0
Ref [8] This work Ref [8] This work Ref [36] This work

Figure 3. 3 Capital, Operating and Overall cost comparison for Case-1, Case-2 and Case-3
respectively with [11] and [43]
60000

5818
50000 7766

Ccapcost
CopD
40000
Cost(Euros)

30000

49259
45658 2765
20000 2704 1671
1066

10000 19163 20775

17599 18067

0
Ref[7] This work Ref[7] This work Ref[7] This work

Figure 3. 4 Capital, Operating and Overall cost comparison for Case-1, Case-2 and Case-3
respectively with [2]

3.9 Sensitivity Analysis

For sensitivity analysis, trials were made by changing electricity cost in the total
cost function in order to check the sensitivity of EA to variation in the economic
parameters. The effect of variation in electricity prices by ±50% with respect to nominal
value was examined. Sensitivity analysis was performed on case-1 data and results are
shown in Table 3.3. These results show that the EA responds correctly by trying to
reduce pressure losses when electricity prices are high at the expense of increased
exchanger surface area, and the opposite is obtained when Cfuel decreases. In case-1,

45
shown in Table 3.3, when Cfuel increased by 50% Capital investment increased by 3.5%
and discounted operating cost CoD increased by 18.8%, but the total cost increased by
only 5.25%. When Cfuel decreased by 50% capital investment decreased by 1% as CopD
decreased by 46% the net cost decreased by 6.3%.
Table 3. 3 Sensitivity to the electricity price (reference exchange in case study # 1)
Cfuel=€0.06 Cfuel=€0.12 Cfuel=€0.18
Ds(m) 0.480 0.692 0.528
B (m) 0.50 0.50 0.50
do (m) 0.015 0.015 0.0179
L (m) 2.896 4.313 3.312
Pt (m) 0.019 0.019 0.022
Cl (m) 0.00375 0.00375 0.00447
Nt 522 1170 438
Vt (m/s) 0.767 0.67 0.6438
Ret 10000 10000 10000
Prt 6.203 5.695 6.203
ht (W/m2K) 10380 9661 8708.14
ft 0.031 0.031 0.031
ΔPt+ ΔPb (kPa) 5.93 6.17 3.89
as (m2) 0.04795 0.06916 0.05279
De (m) 0.011 0.011 0.013
vs (m/s) 0.463 0.536 0.420
Res 6241 12825 6758
Prs 5.394 5.11 5.398
hs (W/m2K) 4412 1970 3867
fs 0.338 0.295 0.333
ΔPs (kPa) 9.5 18.01
U (W/m2K) 1300.16 775.72 1202.33
Sa (m2) 71.14 237.72 76.94
Ccapcost (€) 20,563 45,658 21,491
Cop (€/year) 221 1149 486
CopD (€) 1492 7766 3284
CTOT (€) 22,055 53,425 24,774

It was also attempted to minimize the difference between the shell side and tube
side heat transfer coefficients which is sometimes the criteria for better heat exchanger as
far as the thermal design is concerned. For that case the following objective function is
used
Δh  h s  h t (3.34)
EA is very sensitive to the objective function as expected. The results shown in
Table 3.4, that by taking Δh as an objective function is not productive from the economic
standpoint. The EA was able to make a negligible difference in hot and cold heat transfer
coefficients but at a very high heat exchanger surface area and hence at much higher cost
compared to the original design. Additional capital investment cost incurred will not be
recovered by the reduced operating cost.

46
Table 3. 4 Results for a different objective function (reference exchanger is case study # 1)
Literature values OBJ = Δh
Run 1 Run 2 Run 3
Ds(m) 0.387 0.929 .704 0.722
B (m) 0.305 4.756 3.443 3.526
do (m) 0.019 0.18 0.144 0.147
L (m) 4.880 0.0509 0.0306 0.0320
Pt (m) 0.023 0.064 .038 0.04
Cl (m) 0.004 0.01273 .00766 0.00801
Nt 160 151 252 241
Vt (m/s) 1.76 0.2296 0.3815 0.3635
Ret 36400 10159 10152 10117
Prt 6.2 6.203 6.203 6.203
ht (W/m2K) 6558 3097 5147 4907
ft 0.023 0.031 0.031 0.031
ΔPt+ ΔPb (kPa) 62.812 0.22 1 0.90
as (m2) 0.0236 0.03356 0.02021 0.02125
De (m) 0.013 0.037 0.022 0.023
vs (m/s) 0.94 0.661 1.098 1.044
Res 16200 30275 30246 30087
Prs 5.4 5.394 5.394 5.394
hs (W/m2K) 5735 3097 5147 4907
fs 0.337 0.257 .251 0.251
ΔPs (kPa) 67.684 37.26 118 105
U (W/m2K) 1471 804.3 1110.5 1080.4
Sa (m2) 46.6 115 83.3 85.6
Ccapcost (€) 16,549 27448 22501 22868
Cop (€/year) 4466 800 2784 2474
CopD (€) 27,440 5948 1814 16724
CTOT (€) 43,989 33396 41315 39592

3.10 Exergetic Analysis

Combination of exergy concepts with principles of engineering economy

to optimize thermal systems design is known as thermoeconomic analysis. In this
analysis the real cost sources at the component level are allowed. The available energy
destruction or the exergy losses are resulted from the irreversibilities in the real processes
is the most important part in the exergetic analysis of thermal systems.
In order to carry out the thermoeconomic analysis the exergy change for all flow
streams due to the flow processes must be calculated. For a shell and tube heat
exchanger, exergy changes for hot and cold fluids have to be calculated.
Exergy change due to temperature and pressure is calculated using Eq(3.35) and
Eq(3.36).
  T  C  C  TCin  
 ) min ln 1  ε Hin  1  p, H ln 1  p,C
ψ ΔT  To (C P m 1    (3.35)
   TCin  C p,C  C p, H  THin  

47
 ΔPH ΔP
ψ ΔP  m
H m
C C (3.36)
ρH ρC
Program for exergetic analysis is made according to the Fig (3.10). In this case
the exergy destruction cost is also included. This cost is important for energy efficient
systems. Exergy destruction cost might not be important from manufacturer point of
view but it is important for the efficient use of energy, and for environment protection.
Due to acute shortage of energy it is important to consider that cost of energy which can
be utilized for useful purposes but it is going to be wasted. This is not only a loss of
energy but also a damage of the environment. ψΔT and ψΔP are the thermal and pressure
component of exergy destruction rate.

Capital cost of the heat exchanger is based on the surface area of the heat
exchanger and it is estimated using the Hall method Taal et.al. [45]. suggested that the
Hall method is among the most accurate and simple method to calculate the capital cost.
The equations for estimating the capital and exergy cost are given in Eq(3.37) to
Eq(3.38).

CCHex = 10000 + 324A0.91 (3.37)

Where A is the heat transfer area of the exchanger for given heat duty.

AC  Hr  Cfuel (ψΔT  ψΔP )  FcrCC Hex (3.38)

i(1  i )TL
Fcr 
(1  i )TL  1 (3.39)

Hr, Cfuel, i, TL and Fcr are the time period of operation per year, unit cost of
exergy, interest rate, technical life and the capital recovery factor respectively.

48
Figure 3. 5 Program flow chart for sizing and rating problem and cost optimization

49
3.11 Results and discussions based on exergetic analysis

Some changes are made in the existing MATLAB® program to do the

optimization for the exergetic analysis. In this annual cost of heat exchanger is
optimized. In this optimization problem hot fluid inlet and outlet temperatures (THin,
THout), cold fluid inlet temperature (TCin), hot fluid mass flow rate ( m H ) and tube layout
are specified. In this case annual cost of heat exchanger is optimized including the
exergy destruction cost.

3.11.1 Case study 4: The optimization of shell and tube

In this case study the results are compared with the work of Y.Ozcelik [3]. He
used the Genetic Algorithm to optimize the cost of STHE. In his study, cold fluid mass
flow rate, ratio of baffle spacing to the shell inside diameter and pitch ratio were
considered as varying parameters. He studied the variation of annual cost (AC) vs baffle
spacing ration (BSR) for different tube arrangements and pitch ratio (PR).

In the present study evolutionary algorithm is used to optimize the STHE and
tested. The cold fluid mass flow rate and the ratio of baffle spacing to shell inside
diameter were considered as varying. The discrete variables are standard lengths, tubes
outside diameter, type (square or triangular), pitch ratio, layout angle of the pitch and
number of tube passes.

For the case study, the input parameters are shown in Table-3.5. Some additional
information required for the case is given in Table-3.6. These values are supplied to the
evolutionary optimization algorithm and the optimized parameters for cold fluid mass
flow rate ( m C ) pitch ratio (PR) and baffle spacing ratio (BSR) are obtained for different
discrete variables are shown in Table -3.7 and Table-3.8. The results are compared with
the published results of Y. Ozcelik [3].

50
Table 3. 5 Data for the streams of STHE for case study-4

Fluid m (kg/s) Tin (oC) Tout (oC) Cp (J/kg K) k (W/m K) μ (kg/ms) ρ (kg/m3) Rf (m2K/W)
Tube side - 27 - 4179 0.613 8.55x10-4 997 2x10-4
Shell side 15 77 56 4195 0.668 3.65x10-4 973.7 2x10-4

For the optimization, the range of continuous variable is given below.

10 ≤ m C ≤ 15

1.25 ≤ PR ≤ 1.5

0.2 ≤ BSR ≤ 1 (3.40)

Table-3.6 and Table-3.7 clearly shows thermal component of exergy destruction

is dominant factor on exergy destruction and total cost of the heat exchanger.

Table 3. 6 Data for calculating computer program for case study-4

Hr (h/y) Cfuel ($/W) IR TL (y)
5000 0.00005 0.18 20

The comparison of annual exergy destruction cost and total annual cost for
different tube passes and tube arrangement have shown in Fig-3.11. This is compared
with Ref.[3].

3.11.2 Case study-5: A parametric study in the optimization of a shell and tube
heat exchanger

To observe the effect of the values of continuous variables on the total annual cost
of the heat exchanger. First the optimum values of the discrete variables Table-3.6 were
estimated and fixed during the entire algorithm. The variation of annual cost (AC) vs
baffle spacing ratio (BSR) for different tube layout and different pitch ratio were plotted
and trend is analyzed. Fig-3.11 and Fig-3.12 shows this trend. The trend shown in Fig-
3.11 is that as baffle spacing ratio increases the AC decreases to an optimum point
because the pressure drop decrease with the increase of baffle spacing and hence the
exergy destruction due to pressure drop reduces. The value of baffle spacing ratio in
Table-3.9 (BSR=0.72) for one shell and four tube passes (1-4) shows that algorithm

51
estimate the optimum value for BSR for fixed values of discrete variable. It is clear from
the Fig-3.12 that annual cost is minimum at a particular pitch ratio which is 1.25 in this
study. The data for the Fig-3.12 is shown in Table-3.8. This is in accordance with the
study of optimum pitch for the economic optimization Caputo [2].

Table 3. 7 Design parameter determine by evolutionary algorithm compared with [3]

Do
NSTP m C kg/s L (m)
(mm)
PTT PR PLA BSR

Ref[3] Present Ref[3] Ref[3] Ref[3] Present Ref[3 ] Present

study study study
1-1 10 10 4 8 Square 1.35 1.35 90 0.98 1.0
1-2 10 10 6 12.1 Square 1.30 1.30 90 0.93 1.0
1-4 10 10 7 20 Square 1.25 1.25 90 0.72 0.98
1-6 10 10 5 18 Square 1.25 1.25 45 0.63 0.62
1-8 10 10 3 14 Square 1.25 1.25 90 0.57 0.51

Other parameters like hot and convective heat transfer coefficient, surface area, pressure
drops, and annual and exergy destruction costs are listed with comparison [3] in Table-
3.8 and Table-3.9.

Table 3. 8 The design parameters comparison

A (m2) TC,out (oC) Nt hHot (W/m2-K) hCold (W/m2-K)
Ref[3] Present Ref[3] Present Ref[3] Present Ref[3] Present Ref[3] Present
study study study study study
55 76.7 59 58.6 544 437 8160 9117 4967 3891
68 91.2 59 58.6 299 308 6908 6349 4859 12521
66 95.1 59 58.6 150 146 5819 4222 5257 16639
63 95.6 59 58.6 223 306 5824 4619 6686 15196
62 93.5 59 58.6 473 733 5969 5056 8434 14751

Table 3. 9 : Different component of cost comparison

ΔPH ΔPC ACC ($) ECΔT ($) ECΔP ($)
Ref[3] Present Ref[3] Present Ref[3] Present Ref[3] Present Ref[3] Present
study study study study study
0.277 0.984 0.133 0.059 4179 5010 22076 17384 140 494
0.296 0.486 0.257 0.152 4695 5547 22076 17384 178 288
0.301 0.241 0.530 0.404 4606 5689 22076 17384 249 272
0.264 0.247 1.153 0.395 4494 5707 22076 17384 391 277
0.177 0.226 2.106 0.375 4472 5632 22076 17384 596 265

Table 3. 10 Determined design parameters

NTP L(m) Do(mm) PTT PR PLA m C
4 7 20 Square 1.25 90 10
4 7 20 Triangular 1.25 90 10

52
 Table 3. 11 Determined design parameters
NTP L(m) Do(mm) PTT PR PLA m C
4 7 20 Square 1.25 90 10
4 7 20 Square 1.3 90 10
4 7 20 Square 1.35 90 10
40000

38000

Square Pitch
36000
Triangular Pitch
Square Pitch Ref [13]
34000
Triangular Pitch Ref[13]

32000
AC ($)

30000

28000

26000

24000

22000
0 0.2 0.4 0.6 0.8 1 1.2
BSR

Figure 3. 6 Comparison of AC-BSR for different tube arrangements

34000

32000 PR=1.25
PR=1.30
PR=1.35
PR=1.25 Ref[13]
30000 PR=1.30 Ref[13]
PR=1.35 Ref[13]
AC ($)

28000

26000

24000

22000
0 0.2 0.4 0.6 0.8 1 1.2

BSR
Figure 3. 7 . Comparison of AC-BSR for different pitch ratio

Comparison of total annual cost (Annual Capital + Annual Exergy destruction

cost) for the data shown in Table-3.10 and Table-3.11 is shown in Fig-3.8. Fig-3.8 gives
53
the comparison of this work with the Ref. [3]. Although there is some increase in the
capital cost but the total annual cost is low due to small pressure drops both on shell side
and tube side. The convection heat transfer coefficient on the tube side is higher because
in this study tube side coefficient is calculated according to flow regimes using Eq.3.7,
3.8 and 3.9 mentioned in [11,43]. In all analyzed combinations mentioned in Table-3.6,
the exergy destruction cost reduces. Annual cost of heat exchanger on average is reduced
by 13.5% as compared to the published results in literature [3]. Although there is some
increase in capital cost due to increase in area but this increases the convective heat
transfer coefficient on tube side and hence the exergy destruction due to temperature is
reduced. In total annual cost, the dominant factor is the exergy destruction cost, which is
about 78% of the total annual cost. This confirms that the method represented in this
paper is effective for the cost optimization of a shell and tube heat exchanger.
30000
AC ($)
1-2 1-4 ACC ($)
1-1
1-8
1-6
25000

20000
Cost ($/Year)

22254 22324 22466 22672

15000 22216
17673 17658 17662 17650
17880

10000

5000

5010 5547 5688 5707 5632

4179 4695 4606 4494 4472

0
Ref [13 ] This work Ref [13 ] This work Ref [13 ] This work Ref [13 ] This work Ref [13 ] This work

Figure 3. 8 Comparison of annual capital cost and exergy destruction cost

Shell and tube heat exchanger is designed and optimized the cost for the life cycle
of heat exchanger considering only pumping cost as the component of running cost in the
first phase and then considering the exergetic cost as another component of running cost.
The optimization technique used for cost optimization of the heat exchanger is the
evolutionary strategies (ES). Program is written in MATLAB® environment results show
a significant reduction in the life cycle cost of heat exchanger. A model program is made
for the design and cost optimization of shell and tube heat exchanger using evolutionary
strategies which are best suited for engineering applications.
54
3.12 Conclusion
In first part cases were analyzed taking only pumping cost as the operating cost
for the whole life of the heat exchanger. It is observed that the operating cost is the
dominant factor [71]. In analyzed cases, the operating costs were drastically reduced and
significant life cycle cost reductions were obtained with respect to the original design.
Although there was an increase in capital cost in one of the case but it is completely
offset by the reduction in operating cost. The variation in capital cost ranged between -
11.4% and 25.5% while a percent decrease of operating cost from -45.6% to -90.8% was
obtained. This resulted in the net decrease in total cost saving between -17.1% and -
46.5%, hence confirming the effectiveness of the proposed approach. The result are also
compared with literature values in Caputo [2] and the net cost in two of the cases is less
and in one case there is a small increase. These results indicate that in general
optimization of heat exchanger costs should be searched by acting on the pressure drops,
through proper choice of fluids velocity, rather than minimization of heat exchanger
surface area. Capital investment is concerned with heat exchanger surface area, which is
linked with overall heat transfer coefficient, which is very little effected with flow
velocity ( it depends approximately from Re0.55 on shell side) and can be hardly improved
by changing the design parameters. On the other hand pumping losses are highly
dependent on flow velocity (being depended on Re1.8), hence leaving more space for
design optimization. Other than the cost optimization, the reduction of pumping losses
supports efforts in resource conservation and energy saving which is the main concern on
plant owners.

In second part two cases are analyzed based on exergetic analysis. These two
cases show that EA gives better results as compared with the literature values. The effect
of baffle spacing ratio (BSR) with different tube arrangement on total annual cost is
analyzed and compared with the literature values [3]. This shows that with the increase
in BSR the total annual cost initially decreases rapidly up to an optimum point and then
increases. The effect of baffle spacing ratio with different pitch ratio (PR) on total annual
cost is also performed. This shows that minimum value of total annual cost comes at an
optimum value of PR=1.25.

55
 The proposed method uses the evolutionary algorithm and results of test
problems shows that there is a decrease of 13.5% in annual cost as compared to the
exergetic optimization mentioned in study [3]. So finally, an optimum shell and tube heat
exchanger is obtained and parametric study of continuous variables is done for the user.
The parametric results are comparable with the published results [3].

56
Chapter-4 Optimization of Air-Conditioning Energy Recovery System
using Plate Fin Heat Exchanger

Plate-fin heat exchangers are from the category of compact heat exchanger
designed to pack a high heat transfer capacity into a small volume. They have corrugated
fins or spacers sandwiched between parallel plates (parting sheets). This provides a large
extended heat transfer surface. A view of plate fin heat exchanger is shown in Fig-4.1
and in Fig-4.2
Plate fins are categorized as
1. Straight and plain (i.e. uncut) fins, such as plain triangular and rectangular
fins.
2. Plain but wavy fins (wavy in the main fluid flow direction) and
3. Interrupted fins such as louver, offset strip, and perforated.

Figure 4.1 Basic construction of Plate fin Heat Exchanger Process Heat Transfer [44]

57
 Figure 4.2 A view of Cross-flow plate-fin Heat exchanger Process Heat Transfer [44]

Commonly used fins are shown in Fig-4.3. Plate-fin exchangers have been built
with a surface area density of up to about 5900 m2/m3 (1800 ft2/ft3).

Figure 4.3 Plate fin surfaces examples: (a) Rectangular (plain) (b) rectangular (offset) ; (c)
offset strip; (d) triangular (plain) (e) triangular (perforated); (f) louvered;
(g) wavy Process Heat Transfer [44]

58
 Compact heat exchangers of this type were first developed for applications in
which weight and volume were the premium objective, such as in aircraft and other
mobile units. For this reason aluminum was used for a long time the usual material of
construction and this limited their use to a maximum temperature of around 250°C. For
air-conditioning energy recovery system they are suited because here the maximum
temperature is around 50°C. Stainless steel is now common as the material of
construction for compact heat exchanger operating at temperature of up to 800°C.
In this study plate-fin heat exchanger is used for energy recovery from air-
conditioning system (centrally air-conditioning). Both sides fluid is air. One side has
fresh air (hot fluid) and on the other side is the return air (cold fluid). Flow configuration
is cross flow with both fluids unmixed is used. On both sides offset strip fins are used.
Different types of fins experimental data are listed by Kay & London [46]. This data is
used to get the best geometry configuration for the optimal design of plate-fin heat
exchanger for minimum lifecycle cost and maximum net energy recovered.

4.1 Thermal Design of Plate-Fin Heat Exchanger

Sizing problem of plate-fin heat exchanger deals with the determination of

physical size (length, width and height) of a single-phase cross flow exchanger for
specified heat duty and pressure drop. More specifically inputs to the sizing problem are
surface geometries (including their non-dimensional heat transfer and pressure drop
characteristics), inlet and outlet fluid temperatures, fouling factors, fluid flow rates and
pressure drops on each side.
For the solution to this problem, there are four unknowns, two flow rates or
Reynolds number (to determine correct heat transfer coefficients and friction factors) and
two surface areas for the two fluids cross flow exchanger.

4.1.1 Sizing of a Single-Pass Cross Flow Exchanger

For calculating heat exchanger dimensions for the given heat capacity following
steps are followed. In the sizing problem usually effectiveness and pressure drops are
specified and design is made to satisfy these specifications.

59
4.1.1.1 Heat capacity and outlet temperature

Fluid outlet temperature is calculated from the specified heat duty in order to
compute the fluid bulk mean temperature and the fluid thermo-physical properties on
each fluid side.

q  (m
 Cp)(TH,in  TH,out )  m
 Cp(TC, out  TC, in ) (4.1)

In most of the sizing problems of plate fin heat exchanger effective is specified.
Therefore, hot fluid and cold fluid outlet temperatures can be found as follows.

T H,out  TH,in  ε(Cmin /C H )(TH,in  TC,in ) (4.2)

T C,out  TC,in  ε(Cmin /CC )(TH,in  TC,in ) (4.3)

Initially the air Cp values are taken at the given inlet temperatures of hot and cold
fluid and calculate the outlet temperature using Eq.4.2 and Eq.4.3. Dynamic viscosity
(μ), specific heat at constant pressure (Cp), thermal conductivity (k), Prandtl Number (Pr)
and density (ρ) are taken at the mean fluid temperature.

4.1.1.2 Effectiveness and NTU

By calculating C* and ε (effectiveness) determine NTU from the ε-NTU

expression for cross flow (both fluids unmixed). The longitudinal heat conduction, if
any, is ignored in the first iteration since initially the exchanger size is not known. In
case when both fluids are liquids or both fluids are gases, it can be consider that the
design is balanced (i.e. the thermal resistances are distributed approximately equally on
the hot and cold sides).Shah [4]

C min
C*  ….(4.4)
C max

60
  1
ε  1  exp 

   

NTU 0.23 exp  C* NTU 0.78  1  …(4.5)
 C r  
NTU r  NTU f  2NTU ….(4.6)

4.1.1.3 Core mass velocities

For the selected surfaces on each side some initial guess is taken for Colburn and
Fanning friction factor ratio (j/f) for initial approximation of core mass velocities and
Reynolds number. Initial guess for μo is usually between 0.7-0.8.
1/2
 
 
2g η j
G o 
  1  2/3 NTU f  ….(4.7)
  Pr 
  ρ  

GD h
Re 
μ ….(4.8)

Different geometry specific data is taken from the mentioned values in Kay &
London [46]. One of the set of geometry data is shown in Fig-4.4.

61
 Figure 4.4 Surface basic characteristics for an offset strip fin surface (Kays & London [46])

4.1.1.4 Empirical relationship for Colburn and Fanning friction coefficient

When Reynolds number is found, values of Colburn factor (j) and Friction factor
(f) is again calculated by using Eq.4.9 and Eq.4.10 [31].

0.1541 0.1499 0.0678

0.5403 s   δt   δt 
j  0.06522Re  '    
h   ls  s
0.1 ….(4.9)
 5 0.504  s 
0.504 0.456
 δt   δt 
1.055

 1  5.269  10 Re  '     
  h   ls   s  

62
 0.1856 0.3053  0.2659
 0.7422 s   δt   δt 
f  9.6243Re  '    
h   ls  s
0.1 ….(4.10)
 8 4.429  s 
0.920 3.767
 δt   δt 
0.236

 1  7.669  10 Re  '     
  h   ls   s  
Where,

s  Pf  
h'  b  
b  b1

The geometry dependent properties which are listed in [31]

Pf = Fin pitch
b = Plate spacing
δt = Fin metal thickness
s = Pf - δt
h‟ = b - δt

These correlations are valid for 120≤ Re ≤ 104 [31]

4.1.1.5 Hydraulic diameter

Hydraulic diameter for the passage can be found using Eq.4.11.

4A ocell sh ' ls
Dh   (4.11)
A cell /ls s(sls  h ' ls  h ' δt)  sδδ
Where,
ls = Fin length
Dh = Hydraulic diameter

63
4.1.1.6 Heat Transfer coefficients

After knowing the Reynolds number heat transfer coefficients on hot and cold
side fluids are found using the Eq.4.12.

jGCp
h
Pr 2/3 (4.12)

4.1.1.6 Fin efficiency

First „mf‟ (fin efficiency factor) is calculated using Equation (4.13) for both hot
and cold fluids.

 2h'  δt 
mf   1    (4.13)
 kfδ  ls 
b
lH  lC   δt (4.14)
2
Fin efficiency is calculated using Eq.4.15.

tanh(mf l)
ηf  (4.15)
mf l
The overall efficiency is

 A 
η o  1  1  η f  f 
 A (4.16)

4.1.1.7 Overall Heat Transfer coefficient

Overall heat transfer coefficient is calculated using thermal resistance equation.

1 1 1 α /α α /α
   f r  f r (4.17)
U (η o hA) f (η o hA) r (η o hA) r (η o hA) f

Values of αf and αr are defined as follows.

64
 bf βf
αf  (4.18)
b f  b r  2δ w

b rβ r
αr  (4.19)
b f  b r  2δ w
Where,
Total heat transfer area
β (m 2 /m 3 )
Volume between plates

In initial iteration wall thermal resistance is ignored, because initially the

exchanger size is not known. The ratio of surface area on inlet air side (hot side) and
return air side is:

Ar αr

Af αf (4.20)

4.1.1.8 Surface Area, Free Flow Area, and Core dimensions

Knowing the NTU and Cmin the minimum free-flow area on fresh air side (hot air)
is calculated using Eq.4.2

C min
A f  NTU (4.21)
Uf
Air flow length is computed from the definition of hydraulic diameter. From the
specified m and computed G, the minimum free-flow area on the hot side (fresh air) is

m 
A o,f   
 G f (4.22)

The air flow length is computed form the definition of hydraulic diameter.

 D A
L f   h  (4.23)
 4 Ao  f
65
Similarly for the return air (cold side) surface area and length is calculated.

m 
A o, r   
 G r (4.24)

D A
L r   h  (4.25)
 4A o  r
To calculate the core frontal area on each fluid side, first cross-sectional area to
frontal area () is calculated.

αD h
ζ (4.26)
4
Therefore, frontal area on both sides in now calculated.

Ao
A fr 
ζ (4.27)

After calculating the frontal areas on both sides the length L3 is calculated.
Length L3 is approximately same from both frontal area..

Af Af
L3   (4.28)
LH LC

4.1.2 Rating of a Single-Pass Cross Flow Exchanger

Design obtained from the sizing problem should be checked for exit temperatures
to verify the results. In this case inlet temperatures are given and outlet temperatures are
calculated.

4.1.2.1 Surface Geometries properties

First is to determine the surface geometric properties. This includes the minimum
free flow area (Ao), heat transfer surface area (A), flow length (L), hydraulic diameter
(Dh), heat transfer surface area density (β), the ratio of minimum free-flow area to frontal
area (), fin length Lf and fin thickness (δ).
66
Hydraulic diameter:
4  Passage cross  sectional area
Dh  (4.29)
Passage Perimeter
This is for rectangular passage.

Cross-sectional flow area (S):

S  Area of a passage  No. of passage in each row
(4.30)
 No. of rows allocated to stream
Surface Area (A):
The surface area of the passage is

A  Area of a single row  No. of rows allocated to stream 1 (4.31)

Surface area of fins Af:

The surface area of the fins i.e. the part of corrugated material not in contact with
the plates is

h'
Af  A
h's (4.32)

4.1.2.2 Mean temperature and fluid properties

To determine the mean temperature on each fluid side, first the properties are
evaluated at inlet temperatures for the first approximation. Based on these values outlet
temperatures are evaluated, and the mean temperatures are found. After this properties
are evaluated at this mean temperature.

4.1.2.3 Mass velocities, Reynolds number, Colburn Factor, Friction Factor

Mass velocities, Reynolds number, Colburn factor and Fanning friction factor are
evaluated by using Eq.4.7, 4.8, 4.9 and 4.10 respectively, same as in sizing problem.

67
4.1.2.4 Heat transfer coefficients, fin efficiency
Heat transfer coefficients on hot and cold sides and fin efficiency are evaluated in
the same manner as in sizing problem using Equations (4.12) and (4.15).

4.1.2.5 Wall resistance and overall conductance

For the wall resistance determination, the wall conductance area is Aw.

A w  L f L r (2N P  2) (4.33)
Where NP is number of passes

δW
RW  (4.34)
kWAW
The overall conductance is

1 1 1
  RW  (4.35)
UA (η o hA) f (η o hA) r

4.1.2.6 Effectiveness-NTU and effect of longitudinal conduction

For calculating NTU overall conductance is used.

UA
NTU  (4.36)
C min
Eq.4.5 is used to calculate the effectiveness (ε). The value obtained for the
effectiveness is higher than normally used for the cross-flow heat exchanger. This is due
to longitudinal heat conduction. The decrease in effectiveness due to longitudinal heat
conduction is calculated by first calculating the conduction cross-sectional area for
longitudinal heat conduction in the wall.

A k, H  2N P L r δ W (4.37)

A k, C  (2N P  2)Lfδ W (4.38)

68
 Longitudinal conduction parameters on the hot and cold sides are evaluated using
the following equation.

k A 
λ   W k  (4.39)
 Lr 
Where Cm
 CP
(η O hA) f
Other parameters needed for determining Δε are to find the two ratios.
(η O hA) r

λf
and . From interpolation of tabular results listed by Chiou (published in Shah and
λr
Δε
Mueller [47]) appendix-II, is found. Hence
ε

ε actual  ε  Δε (4.40)

4.1.2.7 Exit temperatures of hot and cold fluid

The heat transfer rate Q is calculated using the following equation.

Q  ε(Tf,i  Tr,i )Cmin (4.41)

The outlet temperatures are finally calculated using the following equations.

Q
Tf,O  Tf,i  (4.42)
Cf
Q
Tr,O  Tr,i  (4.43)
Cr
If the calculated values of exit temperatures are not same then some more
iteration will be done and properties are evaluated at new mean temperatures.

69
4.2 Pressure drop

Pressure drop has to be recalculated to confirm the value of pressure drop. For
the pressure drop calculations, first determine the fluid densities at the exchanger inlet
and outlet (ρi, ρo) of each fluid. The mean specific volume on each fluid side is then
computed from:

1 1 1 1 
      (4.44)
 ρ m 2  ρi ρo 

There is a frequent boundary layer interruptions, in the offset strips fin

geometries, the flow is well mixed and is treated having the Reynolds number very large
(Re → ∞). The entrance and exit loss coefficient are obtained from the Shah [47] for
known , Re, and the flow passage entrance geometry.

Before calculating the pressure drop, it is required to correct the values of

isothermal friction factor by taking into account the temperature-dependent properties. It
is required to compute the fluid bulk mean temperature for both hot and cold fluids

(Tin  Tout )
Tm  (4.45)
2
Where Tm is mean temperature.
The thermal resistance on both sides has calculated using the Eq.4.46
.

1
R th  (4.46)
η o hA

Wall resistance can be ignored, due to small values.

Tm  TW
q (4.47)
R th

70
Hence the wall temperature is,

R 
R f,m   f Tr,m
Tw   Rr 
R (4.48)
1 f
Rr
The corrected friction factor will be
mc
T 
f  f CP  W  (4.49)
 Tm 
Value of „m‟ will be taken from Table. 4.1.

Table 4.1 Values of n and m (source: Data from Shah [48])

Fluid Heating Cooling
Gas n=0, mc=1 for 1<Tw/Tm<3 n=0, mc=0.81, for 0.5<Tw/Tm<1
Liquid n=-0.14, mc=0.58 for μW/μm<1 n==0.41,mc=0.54 for μW/μm>1

71
Table 4.2 Entrance and exit pressure loss coefficients for (a) a multiple circular tube Core, (b)
multiple-tube at tube core, (c) multiple square tube core, and (d) multiple triangular tube
core with abrupt contraction and abrupt expansion [48]

Finally the pressure drop is calculated using following equation.

  ρi  L 1 
ΔP 
G2
  
 1  ζ  Kc  2   f ρ i   m  1  ζ  Ke 
2 2
 (4.50)
2gρ i   ρo  rh  ρ  

Where Kc and Ke are contraction and entrance pressure drop coefficients and are
obtained from figure 4.5 Kay & London [46]. With these values of pressure drops, the

72
outlet pressures were recomputed and with the corresponding values of the new outlet
densities, the pressure drops were recalculated.

4.3 Cost function

Cost of heat exchanger is broadly split in capital and operating cost. Different
methods are used for capital cost. In this thesis capital investment Ccapcost is computed as
a function of the exchanger surface area according to the following equations Sunden
[46].

Ccapcost  CA  A n (4.51)

ACC  a * Ccapcost
r (4.52)
Where a
1  (1  r)TL
Where CA is the heat exchanger investment cost per unit surface area and „n‟ is a
constant. Capital cost is the one time cost occurred during manufacturing.

4.3.1 Operating cost

Heat exchanger operating cost consists of pumping cost for hot (fresh air) and
cold fluid (return air) and the exergy destruction cost due to the irreversibilities. Since
the energy used for operation is electricity, the operating cost is related to the electricity
price. This system is used for energy recovery system for air-conditioning; (energy
recovered from return air) which reduces the compressor work. Reduction in compressor
work is the saving in electricity and hence it is the saving in energy.

4.3.1.1 Pumping cost

Total pumping power is the sum of the pumping power required for return and
fresh air.

PP  Pr  Pf (4.53)
Pumping power required for return (cold) and fresh (hot) side can be calculated in
the same way as in Eq.3.31 and 3.32.

73
  r ΔPr
m
PPr 
ρ r ηr (4.54)

 f ΔPf
m
PPf 
ρ f ηf (4.55)

Where, η is the efficiency of the pump. Pumping cost associated per year can be
calculated once we know the number of hours of operation per day, and the cost of
electricity for each kWh.

 
W PYC  WP  N H  N D  C Fuel (4.56)

4.3.1.2 Exergy destruction cost

Exergy destruction cost is associated with irreversibilities in the system. It

includes exergy change due to temperature difference, exergy change due to pressure
drop, and exergy change due to energy recovery system (which is mostly negative).
Exergy change for the hot air coming in (fresh air) calculated using the following
equation.

ψf  m
 f [(hf,out  h f,in )  To (sf,out  s f,in )] (4.57)

Exergy change for the cold fluid leaving is calculated using the following
equation because it is going to be heated and gain only the sensible heat (only the
temperature rise).

ΔPr T 
ψr  m
 r [C pr (Trout  Trin )   To C Pr ln  rout  (4.58)
ρr  Tfin 

ψ d  ψ f  ψ r (4.59)

74
4.3.1.3 Entropy generation units

The number of entropy generation units is defined as Mishra [49]

S gen
NS  (4.60)
C max
Where S gen is the rate of entropy generation.

S gen  ΔS H  ΔS C (4.61)

And

 T P 
ΔS  m
 C p ln out  Rln out  (4.62)
 T P 

4.3.1.4 Exergy change due to energy recovery system

Fresh air entering the system when passed through plate-fin heat exchanger
(energy recovery system) its temperature decreased and eventually the compressor work
reduced. This is the saving of energy. The amount of energy saved is calculated using
Eq.4.69.

Q
WC  f (4.63)
COP
 m
Q  f  C Pf  ΔT (4.64)
f

Eq.4.63 and 4.64 gives the power in kW. When it is multiplied with the hours of
operation and number of days of operation during a year, it gives the exergy saving due
to energy recovery system. Amount of air exchanged depends upon the tonnage and the
application of air-conditioning system. Values used for case study is for a hospital and
Carrier standards [50] are used.

75
  f  C Pf  ΔT
m
WCY   NH  ND (4.65)
COP
Net exergy destruction cost is

ψ Net  ψ d  WC (4.66)

 dy  ψ
ψ  d N H N D (4.67)

ψ 
 NDy  ψd  W (4.68)
CY

Operating cost for the energy recovery system, including inflation effect.

 1 
((CPPY  ψ dcy )  C fuel ) 

 i  1TL 
OPC  (4.69)
1
i 1

Similarly cost of energy recovered over the life span of air-conditioner is found
using the same approach.

 1 
((WCY )  C Fuel ) 
TL 
 i  1 
TER  (4.70)
1
i 1

Where,
i is the inflation rate and
TL is the life of heat exchanger in years.

Hence, the total life cycle cost (LCC) of energy recovery system is

(4.71)

Hence the net energy recovered (NER)

(4.72)

76
4.4 Modeling Verification

To verify the modeling results, the simulation output are compared with the
published results given in Shah [48]. The operating conditions for the comparison are
given in Table-4.3. The geometry parameters used are geometry-E, from Kays & London
[46].
Table 4. 3 Operating conditions of PFHE (input data for the model)
Volume flow rate of hot (return) fluid (m3/s) 3.494
Volume flow rate of cold (fresh) fluid (m3/s) 1.358
Inlet hot temperature (K) 1173.15
Inlet cold temperature (K) 473.15
Inlet pressure (hot side/return side) (kPa) 160
Inlet pressure (cold side/fresh side) (kPa) 200
Price per unit Area ($/m2) 100
Exponent of nonlinear increase with area increase (n) 0.6
Hours of operation per year (h/yr) 800
Price of electrical energy ($/MWh) 25
Interest rate (%) 14
Life of the heat exchanger (yr) 14
Compressor efficiency 0.65

Comparison of results with the published results of Shah [48] and with Ahmadi
[50].with the same input values given in Table-4.3. Percentage difference of the results
obtained from the modeling and the published results are within that acceptable range.
The comparison is shown in Table-4.4.

77
Table 4. 4 The comparison of modeling output and the published results of Shah[48] and Ahmedi
[50]
Difference
Difference
Output Ref Ref Present (%)
Unit (%) with
Variables Shah[48] Ahmadi[50] study with Ref
Ref[48]
Ahmedi[50]
jH 0.017 0.01749 2.88 0.0168 1.176
fH 0.0669 0.0684 2.242 0.0638 4.634
jC 0.0134 0.0149 11.19 0.0144 -7.463
fC 0.0534 0.055 2.996 0.0521 2.434
hH W/m- 370.5 2.68
360.83 362.68 -0.513
K
hC W/m- 371.4 10.3
336.81 364.122 -8.109
K
Ε 0.8381 0.8444 0.75 0.8273 1.288
Ctotal $ 1621.61 1530.82 0.79 1644.6 -1.418
NS 0.1304 0.1345 3.14 0.1577 -20.935
Δptotal kPa 17.425 16.84 -3.357 17.1469 1.596
Atotal m
2
169.208 174.3 3.01 169.2157 -0.00455

4.5 Energy Recovery System

Due to global warming the use of air-conditioning system has increased

substantially. Air-conditioning systems are used in hospitals, factories, universities,
offices, auditoriums etc. These systems take in fresh air at regular interval of time to
maintain the level of oxygen in the air-conditioned space and to replace the already
conditioned air, which perhaps contaminated by odors, gases, or smoke. As a result of
this ventilation process fresh/hot air goes into air-conditioning space while return/cold air
leaves it. Due to temperature difference between the hot/fresh air goes into and the
cold/return air that leaves the air-conditioned space there would be an unutilized amount
of energy that should be recovered to minimize the energy cost.
Installation of energy recovery systems can considerably reduce the amount of
unutilized energy and can substantially reduce the mechanical heating and cooling
requirements associated with conditioning ventilation air. Moreover, using energy

78
recovery systems can considerably downsize the heating and cooling systems, for which
they reduce peak heating and cooling requirements. In this study energy recovery system
is optimized for the life cycle of heat exchanger including exergy destruction cost.
Payback period of the system is also calculated. The effect of the size of air-conditioning
system on the net energy recovered is also studied.

4.5.1 Types of Energy Recovery System

There are two types of energy recovery systems.

 Heat recovery (HRV) and
 Energy recovery (or enthalpy recovery) (ERV).
Both types include a heat exchanger, one or more fans to push air through the
machine, and some controls. Most energy recovery ventilation systems can recover about
70%-80% of the energy in the exiting air and deliver that energy to the incoming air.
However, they are most cost effective in climates with extreme winters or summers, and
where fuel costs are high.
Energy recovery system used for air-conditioning (centrally air-conditioning) uses
a compact heat exchanger. Fluids on both sides of that heat exchanger is air, which has
low specific heat, hence a large surface to volume ratio (β) is required to get maximum
energy recovery. For ta gas to gas heat exchanger is required. Plate fin heat exchanger is
the type of heat exchanger used for air to air (gas to gas) heat exchange.

4.6 Optimization Methodology

Design of optimum heat exchanger includes the following steps.

i. Based on required amount of heat transfer, we estimate the surface area required
and other design specifications, assuming some initial values.
ii. After this capital cost, operating cost, total cost and amount of saved energy is
calculated
iii. Difference of total energy recovered and operating cost which gives net energy
recovered (objective function).
iv. Using the optimization algorithm (evolutionary algorithm) to select the new set of
values for the design parameters.

79
 v. Repeat the process for different geometries (about 21 geometries) mention in Kay
& London [45].
vi. Select the geometry which gives maximum net energy recovered.
vii. Once a particular geometry is selected optimize the net energy recovered for
different air-conditioning capacities.

Design parameters of energy recovery system are the mass flow rates of fresh and
return air, their inlet temperatures, return air relative humidity and type of geometry (mf,
mr, Tfi, Tfo, Rhr, Geom). Other parameters are found using energy balance. Fixed
parameters of geometry are assigned directly.
The optimization parameters whose values iterated are fresh air (hot air) pressure
drop (δPf), return (return air) air pressure drop (δPr), relative humidity of fresh air (Rhf),
fresh air side length (Lft) and return air side length (Lrt). This process of optimization is
shown in Fig.4.6. Hence the optimization procedure intends to find out the maximum
value of net energy recovered which is the objective function.
Objective function:
Maximize:
NER (total energy recovered)
Subject to:
a. Different geometries with different fin pitch Kays and London [45]
b. m r = the specified value of return air
c. Tf,in = Tfinmin ≤ Tf,in ≤ Tf,inmax
d. Tf.out = fresh air specified value
e. Tr.in = return air specified value
f. Rel-Humdity of Fresh air = Relhumfmin ≤ Reslhumf ≤Relhumfmax
g. ΔPH=ΔPC= ΔPmin ≤ ΔP ≤ ΔPmax

80
 Figure 4. 5 Proposed EA optimization flow chart
This study carried out for different central air-conditioning capacities varies from
50 ton of air-conditioning to 500 ton of air-conditioning. The study gives the type of
geometry, fresh air inlet and outlet temperatures and relative humidity value for which
the lifecycle net saved energy is maximized. Program is made in MATLAB ® for
evolutionary optimization. The flow chart of the program is shown in Fig-4.6.

81
Figure 4.6 Program flow chart for sizing and rating problem and maximization to cost of energy
recover

82
4.7 Results
Optimization is done using evolutionary technique described in Chapter-2.

Table 4. 5 Input data for the case study

Cp k
Volume Flow T input Relative μ
ρ (kg/m3) (kJ/kg (W/m
Rate (m3/s) (ºC) Humidity (Pa s)
K) K)
Case Data:
Fresh Air Mean Mean Mean Mean
3.16 33.9 35-85%
Data: moist air Temp Temp Temp Temp
Return Air Mean Mean Mean
Mean
Data: 35.31 23.9 50% Temp Temp Temp
Temp
moist air

4.7.1 Case study analysis

Data used for analysis and optimization for energy recovery system is of a moderate
size hospital with four operation theaters, twenty private rooms sharing by two patients,
patient wards, a laboratory and other (mess, kitchen etc). Fresh air requirement for each
is taken using Carrier hand book [51]. For such amount of load, air-conditioner load
required is around 100 tons.

Fresh air requirement for such a capacity is found to be 3.16 m3/s, which is
approximately 0.0316 m3/s per ton of air-conditioning. Energy is recovered from the
return air (cold air) such that fresh air (hot air) coming in is at low temperature and hence
reduces the compressor work. Plate-fin cross flow, with both fluid unmixed heat
exchanger is used for energy recovery system. Temperature and thermal properties for
the case are shown in Table-4.5.

Different standard geometries named (A-U) [46] are used for fresh and return air
sides. On both sides same type of geometries are used. Twenty one different geometries
are used to find which geometry gives the minimum cost and maximum energy
recovered. The material of geometries is aluminum alloy 2024-t6 with Cu=4.5%,
Mg=1.5%, Mn=0.6%.

83
 For the optimization variables fresh air pressure drop, return air pressure drop,
fresh air humidity, return air length, fresh air length (δPf, δPr, Relhum_fresh, Lrt, Lft) the
following lower and upper bounds were imposed.

0.5 kPa ≤ δPf ≤ 1.5 kPa

0.5 kPa ≤ δPr ≤ 1.5 kPa
35% ≤ Relhum_fresh ≤ 85%
0.8 m ≤ Lrt ≤ 1.5 m
0.8 m ≤ Lft ≤ 1.5 m

Discounted operating cost were computed with N = 10 years, annual discount rate
i = 10%, energy cost Cfuel = 0.12 $/kWh, daily working hours Hd =18 and number of days
in a year the system is used are Dy =180 days.

4.7.2 Effect of different geometries on cost of energy recovered

For uniform cores different plate-fin geometries naming A to U in Kays &

London [46] (A – U) configuration are used on both sides keeping other parameters like
fresh air inlet temperature, return air temperature, fresh and return air humidities and
pressure drops are constant, value of energy recovered and cost associated with it is
evaluated. Net Saved Energy Vs Different Geometries

245000

240000
Net Saved Energy ($)

235000

230000

225000

220000
A B C D E F G H I J K L M N O P Q R S T U
Geometries

Figure 4. 7 Net saved energy value for different geometries of a 100 tons air-conditioning system

84
 It is found that geometry H gives the maximum value of energy recovered over
the whole life of energy recovery system. In calculating the value of energy recovered,
effect of inflation with time is also included. It is evident from Fig-4.8 that maximum
value of energy recovered for 100
Net Saved Energy tons of
Vs Different air-conditioning
Geometries Capacitiesis of geometry –H.
capacity
for different A/C

1400000

1200000

1000000
Net Saved Energy ($)

100 Tonns
800000
200 Tonns
300 Tonns
400 Tonns
600000 500 Tonns

400000

200000

0
A B C D E F G H I J K L M N O P Q R S T U
Geometries ($)

Figure 4. 8 Net saved energy value for different geometries at different air-conditioning capacities

Fig-4.8 shows that for different air-conditioning capacities geometry-H is still the
best choice as for net energy recovered. The amount of energy recovered increases with
the air-conditioning capacity.

Further investigation on cost of energy recovery systems and net amount of energy
recovered value is large compared to system capital and running cost. Figure-4.9 shows
the same for 100 ton of air-conditioning.

85
 Capital Cost, exergy Destruction Cost, pumping Cost and Net saved Energy Vs different Geometries

240000

Destruction Cost and Net Saved Energy ($)

Capital Cost, Pumping Cost, Exergy
200000

Capital Cost
160000 Pumping Cost
Exergy Destruction Cost
Net Saved Energy
120000

80000

40000

0
A B C D E F G H I J K L M N O P Q R S T U
Geometries

Figure 4. 9 Capital cost, Pumping cost, Exergy destruction cost, and net recovered energy for
different geometries at 100 tons of air-conditioning

Investigation on cost of energy recovery systems and net amount of energy

recovered value shows that for the longer period of time net energy recovered is large
compared to system capital and running cost. Fig-4.9 shows the same for 100 ton of air-
conditioning. It is clear from Fig-4.9 that capital cost, pumping cost and exergy
destruction cost are very small as compared to amount of net saved energy. Exergy
destruction cost is lower than the pumping cost, because in case of air-conditioning
exergy change due to temperature is not high due to low temperature gradients, but
exergy destruction is significant due to pressure drops. Effect of inlet temperature and
humidity on net saved energy is also studied.

4.7.3 Effect of fresh air inlet temperature on amount of energy recovered

Net recovered energy increases as fresh air inlet temperature increases. It is due
to the fact that higher temperature gradient recovered more energy. Hence energy
recovery system becomes more effective in summer.

86
 THin Vs Net Saved energy

355000

305000

255000
Net Saved Energy ($)

205000

155000

105000

55000
307 308 309 310 311 312 313 314 315 316 317 318
Fresh Air Inlet Temperature (K)

Figure 4. 10 Net energy recovered value for different fresh air inlet temperatures of 100 tons of air-
conditioning

4.7.4 Effect of fresh air inlet relative humidity on amount of energy recovered

Relative humidity is the percentage of water vapors present in air compared with
the saturated water vapors present in air at atmospheric temperatures. Fig-4.11 shows
that net amount of energy recovered increases as the relative humidity increases. It is due
to the reason that latent heat is removed and it quiet high. As humidity increases more
latent heat is removed and hence Fresh
larger amount
air relative humidiaty Vsof energy
Net Saved energy recovered.

235000

215000
Net Saved Energy ($)

195000

175000

155000

135000

115000

95000

75000

55000
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Fresh Air Humidity (%)

Figure 4.11 Comparison of net recovered energy for different fresh air relative humidity

87
 Total cost of energy recovery system increases as the humidity increases. It is
due to increase in operating cost [72]. Within the operating cost exergy destruction cost
increases as relative humidity increases, but the total amount of saved energy is higher
compared to increase in operating cost. Hence in higher humidity environment energy
recovered is high, which is feasible in higher humidity season as in July, August and
September months of a year.

4.8 Optimization process

After parametric study it is concluded that geometry-H will give the maximum net
energy recovered for the life cycle of energy recovery system for different refrigeration
capacity.
Optimization results obtained from evolutionary algorithms are confirmed using
Direct Search and Genetic optimization methods which are inbuilt algorithm in EES®
professional version.

For the purpose of this study, is to maximize net energy recovered (NER). For
this 1/NER is minimized. The return air (cold fluid) temperature is taken as 298.2K, its
humidity is taken as 50% (the desired maintained humidity). Fresh air (hot fluid) inlet
temperature is varied from 35°C to 45°C (308.2K – 318.2K) and humidity varied from
35% to 85. Violations of constraints are avoided by adding penalty terms.

1
Minimize:
NER
T fin  308.2 K  318.2 K
Fresh air rel  humidity  35%  85%
Pf  0.5 kPa 1.5 kPa
Pr  0.5 kPa  1.5 kPa

With these constraints and using the evolutionary optimization we achieve the
convergence for different parameters.

88
 Fig-E-1 illustrates the convergence of the best fitness values with respect to the
generations. It is clear that as generations pass, the best fitness values of the individuals
minimized and eventually converges to a value. The problem is to optimize for
minimization hence 1/NER is minimized which gives the maximum of NER.

Fig-E-2, E-3 and E-4 illustrate the convergence of the individuals (δPf, δPr,
rel_humidityf,) to the optimal value. Each individual indicated by a cross. As illustrated
in the figure, the individuals have scattered throughout the entire design space for the first
five generations, after which they start to converge and reach the optimal value.

The optimal values of δPf, δPr, rel-humidityf are determined using the
evolutionary procedure and are found to be 0.5kPa 0.5 kPa, and 85% respectively. These
values are calculated for the 100 ton of air-conditioning load with volume flow rate
3.16m3/s. Maximum net energyNetrecovered value
Saved Energy per tonn is A/C
Vs different $221,525
Capacities or 2215 $/Ton.
2490

2480

2470
Value of Energy Saved/Tonn ($)

2460

2450

2440

2430

2420

2410

2400

2390
0 100 200 300 400 500 600
A/C Capacity (Tonns of Refrigeration)

Figure 4.12 Net energy recovered per ton of air-conditioning for different air-conditioning
capacities at optimum conditions
Optimization carried out for different air-conditioning capacities (50 Tons to 500
Tons) and compared per ton energy recovered for different capacities. Fig- 4.12 shows
that per ton energy recovered at different air-conditioning capacities at optimum
parameters. It is clear that per ton net energy recovered is increasing with the increase in
A/C capacity. It is due to the economy of scale.

89
4.9 Payback period of energy recovery system

Payback period is one of the important factors for the recovery of the investment.
This is the measures of time in which thePeriod
Payback investment is recovered.
Vs A/C Capacities

5
Payback Period (Months)

0
50 100 150 200 250 300 350 400 450 500
A/C Capacities (Tonns of A/C)

Figure 4.13 Payback period of energy recovery system for different air-conditioning capacities
Fig-4.13 shows that payback period is in months and it is decreasing with the
increasing capacity due to economy of scale. Net amount of energy recovered along the
life of energy recovery system is plotted and it is observed that within first year the
system become economically feasible. Cash Flow Vs Time line

380000 363778

326402
330000
289026
Net Energy Recovered ($)

280000
251650

230000 214274

176898
180000
139522
130000
102146

80000 64770

27394
30000

-20000 -9982
1 2 3 4 5 6 7 8 9 10 11
Time (Years)

Figure 4.14 Net amount of energy recovered along the life of energy recovery system

90
4.10 Capital recovery rate

Capital Recovery Rate (CRR) is the payment rate per dollar of investment and the
interest. It is found to be 6.14% per year which is lower than the interest rate of 10% and
quiet feasible investment.
0.115
Cost of Conserved Engergy CCE (cents/kWh)

0.110

0.105

0.100

0.095

0.090

0.085

0.080

0.075
0 50 100 150 200 250 300 350 400 450 500
A/C Capacities (Tonns of A/C)

Figure 4.15 Cost of conserved energy at different capacities

As cost of conserved energy can be directly compared to local energy costs CCE
is the better measure of cost effectiveness than the length of payback period. Its value is
calculated for different capacities of A/C and it is in the range of 0.55 – 0.25 ¢/kWh
which is much less than the residential electricity at 12 ¢/kWh. Figure-4.15 shows
clearly that cost of conserved energy is decreasing with the increase in A/C capacity
because large capacity A/C required less cost per ton of air-conditioning.

91
4.11 Sensitivity analysis

Table 4.6 Sensitivity to the electricity price

Cfuel: 0.06 Cfuel: 0.12 Cfuel: 0.18
$/kWh $/kWh $/kWh
PCY ($) 761 1522 2284
TPC ($) 5146 10291 15437
Exdc/y (4) 404 808 1212
Texdc ($) 2730 5459 8188
TOPC ($) 7875 15750 23625
NetsavEnergy ($) 98213 242644 352839
Capcost ($) 9982 9982 9982
Payback Period (months) 11 10 7
CCE (¢/kWh) 1.62 0.84 0.58

Where PCY, TPC, Exdc/Y, TOPC, NetsavEnergy, CCE are pumping cost per
year, total pumping cost, exergy destruction cost per year, total operating cost, net saved
energy and cost of conserved energy.

Sensitivity analysis is done by changing the electricity price in the total cost
function to assess the sensitivity of EA to variation in the economic parameters. The
effect of change in electricity pricey by ±50% with respect to nominal value was
examined. The results are shown in Table-3.4. These results shown describe that the EA
responds correctly by trying to increase the net energy savings with less payback period
with the increase of fuel cost and making opposite when CF decreases. Values shown in
Table-3.4, when CF increased by 50% pumping cost increased by 48% and discounted
operating cost increased by 50%, but the net saved energy (saved energy value minus the
total operating cost) increased by 57% and the payback period reduced to only 7 months.
When cost of fuel (CF) decreased by 50% pumping cost decreased by 50% total operating
cost also decreased by 50% and net saved energy decreased by 57%. Cost of conserved
energy has decreased from 0.84 ¢/kWh to 0.58 ¢/kWh when cost of energy increased
from 12¢/kWh to 18¢/kWh.

In summary a plate fin heat exchanger is designed and optimized for the
maximum energy recovered from an air-conditioning system. The design is made and
verified with the published results. A case study for a hospital is studied. Optimization is

92
done including the life cycle cost of heat exchanger (capital + exergy destruction cost)
and amount of energy recovered from the system. The effect on net energy recovered
because of capacity, cost of conserved energy and payback period of cost is also studied.

4.12 Conclusion

In this chapter a plate-fin heat exchanger is designed and verified with the
published results. In the next part of this chapter plate-fn heat exchanger is used and
analyzed for energy recovery system for an air-conditioning system. In this case cost of
the system included capital and operating cost. Operating cost includes both the pumping
and exergy destruction cost. The plate-fin heat exchanger is finally optimized using
evolutionary algorithm. The optimization results show that energy recovery system is
feasible for using with air-conditioning systems. The payback period is within the one
year. The cost of conserved energy is also well below then the cost per unit of electricity.
The amount of energy conserved is higher for higher capacity of air-conditioning system.
Number of hours of operation per year will also increase the amount of energy conserved
and will reduce the payback period. Finally a relationship between tons of air-
conditioning and saved energy per ton of air-conditioning is obtained. This relation will
give a fair idea to find amount of energy saved for an optimized energy recovery system
to an air-conditioning system.

93
 Chapter-5 Lifecycle Cost Optimization of Waste Heat Recover
System Using Tube-Fin Heat Exchanger

Due to substantial increase in energy prices and shortage of energy, the energy
recovery systems are getting more importance. Specially to recover waste heat from
thermal power plants and in process industries will reduce wastage of energy and
eventually reduce energy cost. Energy recovery systems not only reduce the cost but also
help in improving the environment by reducing the global warming phenomenon.

Tube fin heat exchangers are employed for the exchange of heat between a liquid
and a gas. These heat exchangers have applications in automobile radiator, HVAC
systems and refrigeration. Poor heat transfer of the gas side is improved by using fins on
the periphery of the liquid carrying tubes. In motor vehicles‟ radiator the hot fluid is
moving in tubes and atmospheric air which is the cooling fluid is moving outside across
the radiator. In case of heat recovery from hot gases, hot gas moves outside the heat
exchanger and cold fluid is moving in the tubes. It is cross flow heat exchanger with one
fluid mixed and other unmixed. These types of heat exchangers are the most successful
type of heat exchanger.

5.1 Tube-Fin Heat Exchanger

Tube fin heat exchanger considered, has the cold fluid water inside the tubes and
waste combustion gases (hot fluid) are moving outside. The heat exchanger is a cross
flow type in which one fluid (combustion gases) is mixed and the other (water) is
unmixed. This heat exchanger is used to recover heat from the combustion gases. Water
is to be heated within given temperature ranges. The heat exchanger is first designed
based on the amount of heat recovered and pressure drop. The design parameters are
verified by rating design which is obtained by design process to check the outlet
temperature of water. A view of the annular fin cross flow heat exchanger is shown in
Fig-5.1.

94
 Figure 5. 1A view of annular fin heat exchanger

5.2 Tube-Fin Heat Exchanger Sizing Problem

The annular fin heat exchanger under consideration is a waste heat recovery
radiator in which the cold liquid water is moving inside the radiator tubes while waste hot
combustion gases are moving outside the tubes. The layout of the heat exchanger is
shown in Fig-5.1 while the arrangement of tubes within the exchanger is shown in Fig-
5.2. The shown exchanger shares some common parameters with well known 12
standard geometry exchangers presented in London and Kays [46]. These parameters are
the material of the exchanger (tubes and fins), frontal area, mass flow rate of the two
fluids, inlet and outlet temperatures of the cold fluid (water ) and the inlet temperature of
hot fluid ( combustion gases). These parameters and their values are shown in Table-5.1.
From the given data and based on the energy balance over the heat exchanger assuming
no heat loss from the whole exchanger, the thermal load and outlet gases temperature is
calculated as

qm
 i C pw (TC, o  TC, in ) (5.1)

q
TH,o  TH,in  (5.2)
 o C pg
m

95
 The specific heat of gases (Cp) is taken initially at the given inlet gases
temperature and then the outlet gases temperature is predicted using Eq.5.2. Fluid
properties like density (ρ), specific heat constant pressure (Cp), thermal conductivity (k),
Prandtl Number (Pr) etc are calculated at mean gases temperature.

Figure 5. 2 Side view of heat exchanger to show tube arrangement

The effectiveness of heat exchanger is calculated as:

Q m i Cpg (TC,in  TC,o )

ε 
Qmax m o Cpw (TH,in  TC,in ) (5.3)

The other design parameters which differ between these exchangers and need to
be optimized for minimum life cycle cost are , tube outside diameter (do), fin length (Lf),
fin pitch (Pf), fin thickness (t), center of center distance between tubes in vertical
direction (h), center to center distance between tubes in the depth direction (d). These

96
design parameters will be calculated for the heat exchanger under consideration
according to the design procedure outlined in the following section.
A code is written in MATLAB® for the optimization using evolutionary algorithms
(EAs).
Table 5. 1 Common parameters between present exchanger and 13 standard

Height of exchanger H = 0.335 m

Width of exchanger W = 0.6 m
Thermal conductivity of tubes (Al) k = 237 W/m K
Thermal conductivity of fins (Al) k = 237 W/m K
Inside/Cold Fluid Water
Outside/Hot Fluid Air
m i cold fluid mass flow rate 1 kg/s
m o outside fluid mass flow rate (hot gases) 1.25 kg/s
Water (cold fluid) inlet temperature TC,in = 290 K
Water (cold fluid) outlet temperature TC,o = 370K
Gas (hot fluid) inlet temperature TH,in = 825 K

5.2.1 Design Methodology

The design methodology considered in this thesis is based on ε-NTU approach for
cross flow annular fin heat exchanger with one fluid mixed and one fluid unmixed. The
heat capacity ratio CR is first calculated as

C
CR  min
C (5.4)
max
5.2.1.1 Effectiveness and NTU

For the given heat exchanger with inside fluid (water) has larger capacity Cmax,
because specific heat of water is high and NTU is calculated as Incopera [52]

NTU  (1/CR )ln[C R ln(1  ε)  1] (5.5)

97
5.2.1.2 Fanning and Friction Coefficient

To proceed the calculation of design parameter the Colburn factor, j and Friction
factor f, are to be determined. Both j and f are estimated from their following relations as
a function of Reynolds number [46];

j  C j Re mj (5.6)

f  C f Re mf (5.7)

The above relations for j and f for each standard geometry are plotted as a straight
line on a log-log plot. Fig-5.3 shows these plots for one of these standard geometries.
Below the Fig.5.3 values are listed for some of the geometrical data related to that
particular geometry. Similar charts for the rest of standard geometries are given in Kays
& London [45]. The average values of j and f for each standard geometry can be
approximated from j-Re and f-Re relations by calculating the values of Cj mj, Cf and mf
according to the following procedure:

98
 do= Tube outside diameter =16.38 x 10-3 m
Pfin= Fin pitch = 276 per m
Dh = Flow passage hydraulic diameter = 6.68 x 10-3 m
t = Fin thickness = 0.25 x 10-3 m
ζ = Free flow area/Frontal area = 0.449
α = Heat transfer area/total volume = 269 x 10-3 m-1
Af/A = Fin area/total area = 0.83
Note: Minimum free-flow area is in spaces transverse to flow

Figure 5. 3 Annular fin standard geometry [45]

The calculation of j can be done by calculating the values of Cj and mj as follows

j  C j Re mj (5.8)

Where

lnj 2  lnj1
mj  (5.9)
lnRe 2  lnRe 1

And C j  e bj
Where

99
b j  lnj 2  m j lnRe 2 (5.10)

In the same manner the value of f is calculated as

f  C f Re mf (5.11)

lnf 2  lnf 1
mf 
lnRe 2  lnRe 1 (5.12)

C f  e bf

b f  lnf 2 m f lnRe 2 (5.13)

Re1, Re2, j1 j2, f1 and f2 are the Reynolds numbers, Colburn factors and friction
factors at any two selected points on the corresponding straight line in the chart. The
above procedure is repeated for the 12 geometries.

For annular heat exchanger parameters Cj, Cf, mj and mf are known for the 13
standard geometries. To find the optimized geometry parameters these are also taken as
varying parameters for optimization. The range of these parameters are taken to be
within the limits of the standard geometries (i.e after the calculation of 13 values of Cj,
Cf, mj and mf for each geometry and the range of each parameter is taken between the
minimum and maximum values).

5.2.1.3 Fin Length

Length of fin is calculated using Equation 5.14.

(d f  d o )
Lf  (5.14)
2

5.2.1.4 Inside Surface Area/Outside Surface Area

Ai Inside surface Area of tubes


A 0 (surfacearea of fins)  2  Bare surface area of tubes

100
Ai Di
 (5.15)
A 0 0.5Pf (d f  d 0 )  d 0  Pf t f d 0
2 2

5.2.1.4 Outside and Inside Heat Transfer Coefficient

Outside Reynolds number and convection heat transfer coefficient is calculated

using the following equations

G Dh
Re o  (5.16)
μ
Where Dh is called hydraulic diameter and it is defined as

4xCross  Sectional Area

Dh 
Wetted Perimeter
Where G is:

o
m
G
ζ A fr (5.17)

Afr is the frontal area of heat exchanger and is calculated in terms of height and width as.

A fr  W  H (5.18)

G Cp
h o  jo (5.19)
Pr 2/3
The number of tubes in the direction of height H is calculated using Equation-5.20.

H
N tH  (5.20)
hcc
Where hcc is the center to center tube distance in vertical direction as shown in
Fig-5.2.
Inside Reynolds number is calculated using Eq.5.21.

101
 4m i
Re i  (5.21)
π μ d i N tr
Inside heat transfer coefficient is calculated using the following equation.
If Re ≤ 2300 then

4K i
hi  (5.22)
di
But if Re> 2300 then inside convective heat transfer coefficient is calculated
using Eq.5.23

Nu  K i
hi  (5.23)
di
Where
(f i  0.125)  (Re i  1000)  Pri
Nu  (5.24)
(1  12.7(fi  0.125) 0.5  (Pri  1))
2/3

And

f i  (0.79lnRe i  1.64) 2 (5.25)

5.2.1.5 Overall Heat Transfer Coefficient

After calculating the inside and outside Reynolds numbers, inside and outside
convective heat transfer coefficients are calculated and then the overall heat transfer
coefficient is calculated using following equation
1
 
 1 d0 di 1 
Uo    ln  
h i
A
2ππ
W h o η o,o  (5.26)
 iA 
 o Ao 

Ai
is the ratio of inside and outside surface areas and is calculated using Equation-5.15.
Ao

102
Where W is the width of heat exchanger (the length of tubes) and ηo is the overall fin
efficiency. Fin efficiency is calculated using the following equation.

tanh(mf L f )
ηo  (5.27)
mf L f

Where M is defined as:

h o Per
mf  (5.28)
A o K fin
Where “Per” is the perimeter of fin surface

Per  π(df  d o ) (5.29)

5.2.1.6 Depth and Number of Tubes Required

Finally the depth of heat exchanger and the number of tubes required is really the
actual achievement for the sizing problem of heat exchanger. The volume of heat
exchanger required is calculated using the following equation.

Ao
Vtot  (5.30)
α
Where Ao is required heat transfer surface area on the fin side and it is calculated
by using NTU approach as.

NTU C min
Ao  (5.31)
Uo
Finally the depth of heat exchanger is calculated as

Vtot
dt  (5.32)
A fr
Number of columns and total number of tubes are calculated as following.

103
 dt
N tW  (5.33)
dcc
N t  N tH  N tW (5.34)

5.2.1.7 Inside and Outside Pressure Drops

After calculating volume, number of passes and number of tubes required,

pressure drops on both sides of tubes are calculated. Based on the pressure drop power
requirement is evaluated. The pressure drop is very small or negligible on gases side if
the following condition is satisfied.

ν o,out  f α Vtot ν m 
(1  ζ )(2
 1)   0   0
ν o,in  ζ A ν
fin i 
(5.35)

If the above condition is not satisfied then the pressure drop is calculated using
Equation-4.36 Incopera [52].

 ν 0,out  f αV ν  ν o,in m i2 

ΔPo  (1  ζ 2 )(  1)   0 tot m  
ν o,in   2ζ 2 A 2  (5.36)
  ζA fr ν o,in  fr 

Tube side or inside pressure drop is calculated by the following equation

8m i2 WN tr f i
ΔPi  2 (5.37)
π (d i ) 5 ρ i Ntp 2

Finally power requirement on both fin side and tube is calculated using the
appropriate pressure drops.

104
  ΔP
 m
W
ρ
(5.38)

Where ΔP  ΔPi  ΔPo .

5.3 Objective Function Calculation

Life cycle cost of heat exchanger has two components, capital and operating cost.
Capital cost is includes material cost and manufacturing cost whereas, operating cost
components are pumping and exergy destruction costs. After calculating the capital and
annual operating cost, the life cycle cost is calculated by including the inflation effect.

C LC  C Capcost  C Opt.Cost (5.39)

For capital cost calculation, different methods are used [45]. In this thesis capital
cost is calculated as a function of surface area of the heat exchanger using the following
correlation which has been considered by Vatavuk‟s [45].

CCapcost  231 (Ao  Ai ) 0.693 (5.40)

Pumping cost for both outside and inside fluids is based on fuel cost. Fuel cost in
case of automobile is the cost per liter of gasoline while in case of energy recovery from
hot stream gases is in kWh. Amount of pumping cost depends on number of hours used
during a year. In case of energy recovery from the hot stream of gases it is calculated as
follows.

  N  N C
CPy  W (5.41)
hr day fuel

The other component of operating cost is the exergy destruction cost due to
irreversibilities in the system. Exergy change on fin side and tube side fluid is calculated
in the following way.

105
   To,out   P  
Δψ o  C po (To,out  To,in )  to    C po  ln    R o  ln o,in
 
  .
  (5.42)
  To,in Po,out
    

 ΔPi Ti, out 

Δψ i  C pi (Ti, out  Ti,in )   C pi  to  ln  (5.43)
 ρ i  1000 Ti,in 

Finally the exergy destruction is

ψ d  (m
 i Δψ i  m
 o Δψ o ) (5.44)

Exergy destruction cost is calculated in the same manner as pumping cost is

calculated. For the life cycle cost the effect of inflation is also taken into account by the
present cost estimation.

 1 
(CPy  C dy )    1
 (i  1)
TL
TOPC  
1 (5.45)

i 1

Hence the total lifecycle cost (TLC) is the sum of capital and total operating cost.

5.4 Optimization Methodology

Optimization process of heat exchanger is based on minimizing surface area,

pressure drop and exergy destruction. These parameters represent the three components
of lifecycle cost (i.e. capital cost, pumping cost and exergy destruction cost).

As shown in the above section the life cycle cost calculation of heat exchanger
requires carrying out a comprehensive thermal and hydraulic design procedure. Thermal
design is to calculate the surface area required for a given heat load under the given
operating conditions. Hydraulic design includes the calculation of pressure drops on tube
and fin sides. From this pressure drop power required to run the pump or fan is
calculated. The details of optimization of lifecycle cost will go through the following
steps.

106
  First particular values of geometric parameters are taken within the given range
 Then based on effectiveness of heat exchanger the surface area required for hot
and cold sides is estimated using Eqs (5.1-5.31)
 Pumping power required based on pressure drops on both sides is calculated using
Eqs (5.36-5.38).
 Exergy destruction is calculated next using Eqs (5.42-5.44).
 After this objective function based on capital and operating cost is calculated.
 The optimization algorithm is applied to select new set of values for design
parameters (do, Lf, Pf, t, hcc, dcc, Cj, Cf, mj and mf).
 The design parameters and objective functions for the predefined number of
generations have been iterated till minimum lifecycle cost is obtained.

Design parameters for optimization are different geometry parameters like tube outer
diameter (do), fin diameter (Lf), fin pitch (Pf), fin thickness (t), tube adjacent vertical
height (h), tube adjacent horizontal distance (d). Range of these parameters is taken from
the twelve geometries given in Kays & London [46]. These parameters are iterated to get
the minimum lifecycle cost and particular values of these parameters to define new
geometry which gives minimum lifecycle cost. The optimization process is shown in
Fig.5.4.

107
 Figure 5. 4 Proposed optimization flow chart

Values of hydraulic diameter (dh), free flow area/frontal area (ζ), heat transfer
area/total volume (α) and fin area/total area (Af/At) are calculated based on different
varying parameter of geometry.

dh  4 rh (5.46)

 Ac 
rh  w   (5.47)
 A
Ac  (hcc  d o )(Pf  t ) (5.48)

A
2

π 2

d f  d o2  πd o Pf  t  (5.49)

Ac
ζ (5.50)
A frg
A frg  hcc(Pf  t)  0.9t  (5.51)

4ζ
α (5.52)
dh
108
Af
 2

d f2  d o2 
A  
d f  d o2  2d o (Pf  t)
(5.53)

(d  d o )
Lf  f (5.54)
2

5.4.1 Optimization Problem Description

The aim of optimization is to minimize the lifecycle cost of heat exchanger, under
the following constraints.

Objective function:
To Minimize:
CLC (total lifecycle cost)

Subject to:
m i , m o , TC,in ,TC,o, TH,in have specified values shown in Table-5.1

And the geometric parameters are within the specified range.

a 0.00965≤ do ≤0.02601
b 0.0056≤ Lf ≤0.00905
c 0.0028089≤ Pf ≤0.00362318
d 0.00025≤ t ≤0.00048
e 0.024765 ≤ hcc ≤0.0782066
jf 0.02032 ≤ dcc ≤0.0524
g. 0.22239 ≤ mj ≤ 0.51139
h 0.14189 ≤ mf ≤ 0.28628
i. 0.046346 ≤ Cj ≤ 0.4209
j. 0.08821 ≤ Cf ≤ 0.51614

Evolutionary strategy is used to minimize the lifecycle cost of energy recovery

system. The method of evolutionary strategy is defined in chapter-2. The optimization
procedure is carried out on personal computer (PC) using evolutionary. In this work the
minimum lifecycle cost of heat exchanger is determined by the given limits of varying
parameters define in Kays & London [45]. In the program the number of individuals
assigned to every generation is set to 35 while the maximum number of generations is
selected in the range of 100 to 500. It was observed that the test of convergence was
always obtained within 100 generations.

109
 As an example of population convergence of design parameters, Fig- F-1 shows
the convergence of population with respect to generations. It clearly shows that
convergence is achieved within 150 generations.

5.5 Results and Discussion

In the analysis for cost calculation the cost of energy is taken in electricity cost in
kWh and the rate is taken as $0.12/kWh Caputo [2]. Operating cost with inflation was
computed with N=10 years and annual inflation rate is 10%. The results for cost
optimization are obtained for two cases related to operating cost. In the first case the
results of optimized annular fin heat exchanger are obtained for the case of operating cost
including exergy destruction cost while in the second case the results are obtained
without including the exergy destruction cost. In the first case, obtained results for the
optimized exchanger are presented in Fig-5.5 and in Table-5.2. The results in the second
case are listed in Table-5.3. Fig-5.5 shows the lifecycle cost of the present optimized
exchanger along with the results of the 12 standard geometries. The Fig-5.5 clearly
shows that the lifecycle cost of the new optimized annular fin geometry heat exchanger
has the minimum value for lifecycle cost.
Lifecycle Cost Vs Geometry

460000
455874
455000 452099
449013
450000 447953
444855 444790
445000 442981 443627 443472
442599 442002 441647
Lifecycle Cost ($)

440000

435000

430000
424893
425000

420000

415000

410000

405000
A B C D E F G H I J K L New_Geom
Geometry

Figure 5. 5 Lifecycle cost of waste heat recovery radiator including exergy destruction cost

So the new obtained geometrical design parameters define a new annular fin heat
exchanger with minimum lifecycle cost. These new values of geometrical parameters,
operating cost and total lifecycle cost are listed in Table-5.2. It also shows the operating
cost and total lifecycle cost for two standard geometries of heat exchangers, the first is

110
the standard geometry with identity D which has analyzed in Incopera [52] and the
second is the standard geometry with identity “I” which has the minimum lifecycle cost
among the 12 standard geometries.
Table 5. 2 Comparison of heat exchanger design considering pumping and exergy
destruction costs with true literature values
Case Study Min. Cost Using This Work
Literature values [52] Standard Geom. With Min Cost
Geometry Type D I New Geom.
Ao (m2) 4.0212 5.8386 1.36381
Ai (m2) 1.76 1.297 0.52
Vtot (m3) 0.02 0.0419 0.0123
Ntc 4 6 4
Nt (# of tubes) 57 35 21
Capcost 709 811 347
PPy (kWh) 19189 7383 58
CPPy ($) 2303 886 8
TPPC ($) 15564 5996 55
ψdy (kWh) 532230 536120 523364
Cψy ($) 63868 64334 62804
Texd ($) 431680 434840 424492
CostY 66880 66031 63159
CLC ($) 447953 441647 424893

As declared above the operating cost in Table-5.2 includes both pumping and
exergy destruction cost. The table shows a significant decrease in heat exchanger surface
area which is 11 times less than that of the area calculated in case D Incopera [52]. This
decrease in surface area decreases the pumping cost by $2294 per year as compared to
geometry D and $878 as compared to geometry I (standard geometry which gives
minimum cost amongst 12 geometries). But the exergy destruction cost reduces to only
2% as compared to geometry D. Hence, small decrease in exergy destruction cost is due
to decrease in pumping cost because the temperature limits are same in both cases. This
decreases the overall annual operating cost by 5.15% but the pumping cost reduces to
$15509 as compared to geometry D and $5941 as compared to geometry I. The overall
decrease in lifecycle cost is ($23,060) for new geometry. The large decrease in pumping
cost and capital cost is due to large reduction in surface area of the heat exchanger.
Lifecycle cost in each case is calculated by considering 10% inflationary effect. Detail
parameters are shown in Table 5.2. The comparison of capital, operating and total cost is
shown in Figure 5.6.

111
 1000000

Comparison of Capital and Operating Costs ($)

100000

Operating Cost
447244 Capital Cost
10000 424546

1000

100

709
347
10

1
Literature This Work

Figure 5. 6 Comparison of capital and operating cost (pumping + exergy destruction cost)

Table-5.3 shows the optimized geometry parameters as well as the lifecycle cost
for the case in which the pumping cost is taken as the only component of operating cost
(i.e. excluding the exergy destruction cost but 10% inflationary cost is included). The
cost function in this case is the sum of capital and pumping cost. The lifecycle cost using
new optimized geometry and standard geometry D is shown for the purpose of
comparison. It can be seen from the table that capital cost decreases by 2 times for the
new optimized geometry as compared to geometry D. It also shows that the operating
cost which is the pumping cost is reduced by $15,509 as compared to geometry D. The
annual savings in operating cost is ($2,296) and the overall cost is reduced by $15,864 as
compared with geometry D. Further comparison of capital and operating costs in the two
cases is shown in Fig-4.7.
Table 5. 3 Comparison of heat exchanger design considering capital and pumping cost
Case Study This Work
Literature values [52] With Min Cost
Geometry Type D New Geom.
Ao (m2) 4.0212 1.36381
Ai (m2) 1.76 0.52
Vtot (m3) 0.02 0.0123
Ntc (passes) 4 4
Nt (# of tubes) 57 21
Capcost 709 354
PPy (kWh) 19189 68
CPPy ($) 2303 8
TPPC ($) 15564 55
CostY 3012 362
Lcost ($) 16273 409

112
 100000

10000
Comparison of Capital and Operating Costs ($)

Operating Cost
15564 Capital Cost

1000

55

100

709
347
10

1
Literature This Work

Figure 5. 7 Comparison of capital and operating cost (pumping cost)

The analysis of the results in both cases shows that the capital cost is not the
dominant factor as compared to operating cost. It is seen that the capital cost in both
cases decreased as well as the operating cost. These cases show that the Evolutionary
Algorithm (EA) gives better results as compared to literature values by finding the
optimized geometric parameters for the new geometry. The new geometry not only
decreases the capital cost but reduces the pumping cost to the significant amount. There
is significant decrease in pumping cost but the decrease in exergy destruction cost is not
significant because of same temperature limits on hot and cold fluid. Hence in both cases
(cost with and without exergy destruction) has decreased. Finally, the results obtained
for optimized annular fin geometry has given optimized relations for Colburn and
Friction factors as functions of Re number. These relations read as:

j  0.222390R e 0.4209 (4.46)

f  0.26035R e 0.133083 (4.47)

The above relations are plotted to give chart similar to those given in Kays $ &
London [46] for each of the 12 standard geometry as shown in Fig-2.8. The
corresponding design parameters are listed below.

113
 do= Tube outside diameter =13.25 x 10-3 m
Pfin= Fin pitch = 279 per m
Dh = Flow passage hydraulic diameter = 16.4 x 10-3 m
t = Fin thickness = 0.480 x 10-3 m
ζ = Free flow area/Frontal area = 0.702
α = Heat transfer area/total volume = 172 m-1
Af/A = Fin area/total area = 0.8443
Note: Minimum free-flow area is in spaces transverse to flow

Figure 5. 8 Optimized annular fin geometry

In this chapter a tube-fin heat exchanger is designed and optimized using

evolutionary algorithms (EA) for the life cycle of heat exchanger for waste heat recovery
from waste gases. The aim to recover maximum amount of energy with minimum life
cycle cost of the tube fin heat exchanger. The cost function includes capital cost,
operating cost (exergy destruction cost) and effect of inflation of the cost is also
considered. By taking different standard geometry parameters given in Kays & London
[46] as varying parameters optimal annular fin geometry is found which gives minimum
life cycle cost. The simplified relations for Colburn factor and friction factor are also
determined [73].

114
5.6 Conclusion

In this chapter an optimized design for annular fin heat exchanger that is based on
economic consideration has been carried out. The heat exchanger is selected to have the
same material, frontal area, effectiveness and thermal load of standard geometry heat
exchangers. The optimization procedure for the design parameters has been conducted
using Evolutionary Algorithm (EA). The objective function in the optimization process
was the total lifecycle cost which includes both capital and operating costs. The
operating cost includes pumping and exergy destruction costs plus 10% inflation rate.
The optimization of the lifecycle has been carried out for operating cost with and without
exergy destruction. The optimization of the lifecycle cost for the case of operating cost
without exergy destruction cost has shown significant cost reduction as compared with
that for standard geometry considered in the literature. The analysis of the results has
shown that both capital cost and operating cost decrease as a result of the surface area
reduction of the optimized design which leads to significant reduction of the total life
cycle cost. On the other side relations for both Colburn and friction factors are developed
for the optimized geometry design of the heat exchanger.

115
Chapter-6 Conclusions and Recommendations for future work

In first phase, method of solution of a shell and tube heat exchanger design
optimization problem was proposed based on the utilization of evolutionary algorithm
using Evolutionary Strategy. By doing the analysis on the literature test cases, and
applying evolutionary strategy for optimization a reduction of capital investment up to
11.4% and savings in operating costs up to 90.8% were obtained, with an overall
decrease in life cycle cost up to 46.5%. Furthermore, the evolutionary strategies allow
rapid solution of the design problem and enable to examine number of alternative
solutions of good quality, and hence this gives the designer more degree of freedom in
the final choice as compared to the traditional methods.

Sensitivity analysis in cost due to electricity prices fluctuations is also done, and
this shows higher capital cost in case of high energy prices and less capital cost in case of
low energy prices, which shows the proper working of evolutionary algorithm. The
following important contributions are made in this thesis regarding shell and tube heat
exchanger.

In second part of shell and tube heat exchanger two cases are analyzed based on
exergetic analysis. The results show that that EA gives better result as compared to
literature values [3]. The BSR trend over the annual cost gives the similar trend a given
in [3] and also the pitch ratio (PR) gives the minimum total annual cost at PR=1.25. The
EA gives a decrease of 13.5% in annual cost as compared to the exergetic optimization
mentioned in study [3]. So finally, a better optimum shell and tube heat exchanger has
been designed. In short in this part of the thesis following main contributions are made.

 Shell and tube heat exchanger is optimized based on pumping and well as based
on exergy destruction cost using ES (EA). The results are compared with the
published results.

 The life cycle cost of shell and tube heat exchange is calculated for the technical
life of the heat exchanger including the effect of inflation. Sensitivity analysis is
done on energy prices to check the validity of the software.

116
 In second phase of this thesis, a method of solution for cross flow (both fluid
unmixed) plate-fin heat exchanger design optimization problem for energy recovery
system in a central air-conditioning system was proposed based on the utilization of
evolutionary algorithm. By doing the analysis on test case, cost of capital recovery is
only 0.85 ¢/kWh which is quiet less to the cost of energy (12 ¢/kWh). Payback period is
within the range of one year which is quiet feasible. Hence energy recovery system is
economically viable. It also shows that energy recovery system is more viable for larger
air-conditioning system as compared to small air-conditioning systems. This analysis
proposes the standard geometry listed in Kay & London [46] which gives the minimum
cost of energy recovery system and hence the maximum net energy recovered. The
optimization technique used is the evolutionary strategy and it allows rapid solution of
the design problem. As a conclusion of this part the following important contributions
are made in this thesis.

 For a plate fin heat exchanger an annular fin geometry is selected out of the given
geometries in Kays and London which gives the minimum life cycle cost and
gives the maximum net energy recovered from an A/C system for its life cycle.

 Effect of temperature, humidity is studied on the net energy recovered from the
air-conditioning system..

 CCE (Cost of conserved energy) and payback period of the system is calculated
and its variation with the air-conditioning capacity is observed. It is found that
CCE is well below the cost of electricity per KWh and the payback period is also
less than a year.

 In third phase of the thesis solution methodology of cross flow tube-fin heat
exchanger design (one fluid mixed and one unmixed) and optimization problem
for waste heat recovery system was proposed based on the utilization of
evolutionary algorithm. Capital cost has reduced significantly and also the
pumping cost and exergy destruction cost, which leads to the reduction of life
cycle cost of heat exchanger.

117
 In this optimization a new annular fin geometry parameters are obtained using
evolutionary algorithms which gives the minimum life cycle cost. The new
geometry parameters are optimized within the range of twelve standard
geometries given in Kay & London [46].

 Colburn and friction factor simplified relationship are also proposed for annular
fin geometry.

For future work, study required for detail mechanical design and the heat
exchanger manufacturing process of shell and tube heat exchanger, plate-fin heat
exchanger and tube-fin heat exchanger to significantly improve the estimation capability
of the capital investment to obtain a more realistic heat exchanger optimization process.

Further study can be carried out that if an energy recovery system is designed for
new air-conditioning system, what will be the effect on capital cost of air-conditioning
system.

118
Appendices

119
 Appendix-A

Geometry A: Strip-fin plate-fin surface

120
Geometry B: Strip-fin plate-fin surface

121
Geometry C: Strip-fin plate-fin surface

122
Geometry D: Strip-fin plate-fin surface

123
Geometry E: Strip-fin plate-fin surface

124
Geometry F: Strip-fin plate-fin surface

125
Geometry G: Strip-fin plate-fin surface

126
Geometry H: Strip-fin plate-fin surface

127
Geometry I: Strip-fin plate-fin surface

128
Geometry J: Strip-fin plate-fin surface

129
Geometry K: Strip-fin plate-fin surface

130
Geometry L: Strip-fin plate-fin surface

131
Geometry M: Strip-fin plate-fin surface

132
Geometry N: Strip-fin plate-fin surface

133
Geometry O: Strip-fin plate-fin surface

134
Geometry P: Strip-fin plate-fin surface

135
Geometry Q: Strip-fin plate-fin surface

136
Geometry R: Strip-fin plate-fin surface

137
Geometry S: Strip-fin plate-fin surface

138
Geometry T: Strip-fin plate-fin surface

139
Geometry U: Strip-fin plate-fin surface

140
Appendix-B

141
 Appendix-C
Finned Circular tubes, Surface

Finned Circular tubes, surface A

142
Finned Circular tubes, surface B

143
Finned Circular tubes, surface C

144
Finned Circular tubes, surface D

145
Finned Circular tubes, surface E

146
Finned Circular tubes, surface F,G

147
Finned Circular tubes, surface H,I,J and K

148
Finned Circular tubes, surface Land M

149
 Appendix-D
Shell and Tube Heat Exchanger convergence and variable parameters
D-1
4
x 10
2.75

2.7

2.65
Total cost in Euros

2.6

2.55

2.5

2.45

2.4

2.35
0 50 100 150 200 250 300 350 400 450 500
# of generations

Fig.D-1 Convergence of fitness with respect to generations

150
 0.055

0.05

0.045
Tube outer diameter (m)

0.04

0.035

0.03

0.025

0.02

0.015
0 50 100 150 200 250 300 350 400 450 500
# of generations

Fig.D-2: Convergence of populations with respect to generations of do

0.5

0.45

0.4

0.35
Baffle spacing (m)

0.3

0.25

0.2

0.15

0.1

0.05
0 50 100 150 200 250 300 350 400 450 500
# of generations

Fig.D-3: Convergence of populations with respect to generations of baffle spacing (B)

151
 1.4

1.2

1
Shell diameter (m)

0.8

0.6

0.4

0.2

0
0 50 100 150 200 250 300 350 400 450 500
# of generations

Fig. D-4: Convergence of populations with respect to generations shell diameter ds

152
 Appendix-E
Plate-Fin Heat Exchanger Convergence Figures

Figure E-1: Convergence of fitness with respect of generation

2.5
Individuals deltaPr

1.5

0.5
0 10 20 30 40 50 60 70 80 90 100
Generations
Figure E-2 Convergence of populations with respect to generations of δPr

153
 3

2.5

Individuals deltaPf
2

1.5

0.5
0 10 20 30 40 50 60 70 80 90 100
Generations

Figure E-3: Convergence of populations with respect to generations of δPf

0.9

0.8
f
Individuals relhum

0.7

0.6

0.5

0.4

0 10 20 30 40 50 60 70 80 90 100
Generations

Figure E-5: Convergence of population with respect to generations of fresh air relative humidity

154
 Appendix-F
Tube-Fin Heat Exchanger Convergence Figures
5
x 10
4.2525

4.252

4.2515
Life Cycle Cost ($)

4.251

4.2505

4.25

4.2495

4.249

4.2485
0 50 100 150 200 250 300
Number of Generations
Figure F-1: Convergence of fitness with respect to generation

0.022

0.02

0.018
Individuals do

0.016

0.014

0.012

0.01

0.008
0 50 100 150 200 250 300
Generations

Figure F-2: Convergence of populations with respect to generations of do

155
 -3
x 10
3.8

3.7

3.6

3.5
Individuals Pf

3.4

3.3

3.2

3.1

2.9

2.8
0 50 100 150 200 250 300
Generations

Figure F-3: Convergence of populations with respect to generations of Pf

-4
x 10
5

4.5

4
Individuals t

3.5

0 50 100 150 200 250 300

Generations

Figure F-3: Convergence of populations with respect to generations of t

156
 0.08

0.075

0.07

0.065
Individuals height h

0.06

0.055

0.05

0.045

0.04

0.035

0.03
0 50 100 150 200 250 300
Generations
Figure F-4: Convergence of populations with respect to generations of h

0.042

0.041

0.04

0.039
Individuals d

0.038

0.037

0.036

0.035

0.034
0 50 100 150 200 250 300
Generations

Figure F-5: Convergence of populations with respect to generations of d

157
 0.5

0.45

0.4
Individuals mj

0.35

0.3

0.25

0.2
0 50 100 150 200 250 300
Generations
Figure F-6: Convergence of populations with respect to generations of mj
0.32

0.3

0.28

0.26
Individuals mf

0.24

0.22

0.2

0.18

0.16

0 50 100 150 200 250 300

Generations

Figure F-7: Convergence of populations with respect to generations of mf

158
 0.45

0.4

0.35

0.3
Individuals cf

0.25

0.2

0.15

0.1

0.05
0 50 100 150 200 250 300
Generations

Figure F-8: Convergence of populations with respect to generations of Cf

0.45

0.4

0.35

0.3
Individuals cj

0.25

0.2

0.15

0.1

0.05

0
0 50 100 150 200 250 300
Generations

Figure F-8: Convergence of populations with respect to generations of Cj

159
 Appendix-G

Matlab® Code
Start Programme
close all
clear all
clc
tic %start the stop watch

path(path,'myES')

global display
%load data

display=0;

rand('state',sum(100*clock));
randn('state',sum(100*clock));

runs = 1;

% define optimization task

opttask.dim = 3;
opttask.range = [10 13; % mC (Tube side mass flow rate)
0.2 1; % BSR (baffle spacing ratio)
0.05 1.5]; % ds (shell diameter)
opttask.func = @HXG;
% define ES parameters
% esopt.mue = 5;
% esopt.lambda = 35;
esopt.mue = 5;
esopt.lambda = 80;
esopt.maxGenerations = 100;
esopt.recombination = 'intermedian';
esopt.rho = 2;
esopt.selection = 'plus';

for r = 1:runs
disp(['run: ' num2str(r) ' ' datestr(now)])
clear pop pophistory
[pop,pophistory] = myES(opttask,esopt);
% delete data in each generation exept of the last to save memory
for g = 1:esopt.maxGenerations
pophistory{g}.data = [];
end
end

f=[];
for i=2:esopt.maxGenerations
f=[f pophistory{i}.bestFit ];
end

160
plot(f)

% Plot and Display Results

offsprings_mC=[];
offsprings_BSR=[];
offsprings_ds=[];

for i=2:esopt.maxGenerations,
offsprings_mC=[offsprings_mC pophistory{i}.offspring(1,:)'];
offsprings_BSR=[offsprings_BSR pophistory{i}.offspring(2,:)'];
offsprings_ds=[offsprings_ds pophistory{i}.offspring(3,:)'];
end

figure
plot(offsprings_mC','bx')
xlabel('Generations')
ylabel('Individuals mC');

figure
plot(offsprings_BSR','bx')
xlabel('Generations')
ylabel('Individuals baffle sapcing');

figure
plot(offsprings_ds','bx')
xlabel('Generations')
ylabel('Individuals ds');

figure
plot(f(1,:))
xlabel('Generations')
ylabel('Best Fitness');

disp(sprintf('mC=%1.4f[m]',pophistory{end}.best(1)))
disp(sprintf('BSR=%1.3f[m]',pophistory{end}.best(2)))
disp(sprintf('ds=%1.3f[m]',pophistory{end}.best(3)))

display=1;
HXG(pophistory{end}.best);
ttt=toc

161
Shell and Tube Heat Exchanger Code
TUBE SIDE CALCULATIONS
%********************TUBE SIDE
CALCULATIONS******************************** "
[n1, k1]=ntubes(Tarr,np)
nt=k1*(ds/do)^n1;
"np= number of passes" "Nt= no.of tubes"

At = (3.141/4)*(di^2)*nt; % The total flow area through the tubes"

velc=(4*mCES*np)/(3.141*(di^2)*denC*nt)
ret=(denC*velc*di)/dvisC; %Tube side (cold Fluid) fluid Reynolds
number"
fc=(1.82*log10(ret)-1.64)^(-2)
ht=htt(ret,fc,prtc,kC,di,l)

Overall Heat Transfer Coefficient

%****************** Overall Heat Transfer Coefficient
******************** "
uc=((1/ho)+(do/(di*ht))+(do/2)*(log(do/di)/kw))^(-1);
%Clean overall heat transfer coefficient bases on outer surface area"
uf1=((1/ho)+(do/(di*ht))+Rfs+(do/di)*Rft+(do/2)*(log(do/di)/kw))^(-1);
% Fouled overall heat transfer coefficient based on outer surface area"

%************************ LMTD Calculation ************************ "

dt1 = THin-TCout;
dt2 = THout-TCin;
LMTD = (dt1-dt2)/log(dt1/dt2);

P= (TCout-TCin)/(THin-TCin);
R= (THin-THout)/(TCout-TCin);

F factor function

F=Ffactor(P,R,np) % Call tthe Ffactor function to get the value of F

dtm=F*(LMTD);

ac= (q*1000)/(uc*dtm) ; %The surface area of the heat exchanger for the
clean condition"

uf=uc/1.2; %Let assume over design of 20% and periodic cleaning

schedule will be there"

af1= (q*1000)/(uf1*dtm)

Rft1=(1/uf)-(1/uc); %Total fouling based on 20% extra design"

af=1.2*ac; % 20 over surface design, the surface area of the heat

exchanger"

% ********************* LENGTH OF THE HEAT EXCHANGER

********************** "

% ****************** PRESSURE DROP CALCULATION *********************"

162
% *********** Tube-Side Perssure Drop ******************************"
dpt=((denC*(velc^2)*l*fc*np)/(2*di))*0.001 % Tube-side pressure drop in
kPa
dpb=((denC*(velc^2)*2.5*np)/2)*0.001; % Pressure drop in a bend of 180
degrees"

%************ Shell-Side Pressure Drop ************************* "

fs= exp(0.576-0.19*(log(res)));

Gs= (mH)/(as); % The Shell-side mass velocity"

Tw=(1/2)*((TCin+TCout)/2+(THin+THout)/2);

kinvisw=94.671*exp(-0.0401*Tw); % Kinematic viscosity at wall

temperature"
dvisw = denH*kinvisw; % Dynamic Viscosity at wall temperature"
phs=(dvisH/dvisw)^0.14;
%phs=0.97

nb=abs(floor(-((l/bs)-1))); %Number of baffles"

%
dps=((fs*(Gs^2)*(Nb+1)*Ds)/(2*denH*Dse(Tarr$,pt,do)*phs))*convert(Pa,kP
a);" "Shell side pressure drop"
dps=(denH*(velh^2)*fs*l*ds)/(2*bs*dse*phs)*0.001

% ****************** PUMPING POWER CALCULATIONS ******************* "

% ******************** Tube-side pumping***************************"

effp= 0.8; %Pump efficiency"

Pwt=(mCES*dpt)/(denC*effp); % Tube-side pumping power required"

% Pwte=(mC*dpte)/(denC*effp); "" Tube-side pumping power required"

Pwb=(mCES*dpb)/(denC*effp); % Bend pumping power required"

% ********** Shell-side pumping power based on Sizing ***************"

Pws=(mH*dps)/(denH*effp); % Shell side pumping power required in kW"

% Pwse=(mH*dpse)/(denH*effp);" " Shell side pumping power required in
kW"

Pwtot= Pwt+ Pws; % Total Pumping power required in kW with fouling "
Pwtot1=Pwt+Pws+Pwb ; % Total pumpoing power required in kW including
pressure loss in bends"
%********** Effectiveness using 1st Law of thermodynamics ***********

C min function

Cmin = Cminn(Cpc,CpH,mH,mCES)
Effec=(THin-THout)/(THin-TCin)
to= 300; %Environmental Temperature "

163
% ******** Exergy Change in the Cold Fluid (Tube-Side Fluid)
************"
%excc = mCES*(Cpc*(TCout-TCin)+((dpt+dpb)/denC)-
to*Cpc*log(TCout/TCin)) %exergy change in kW"
%*************** Exergy Change in the HotFluid (Shell-Side Fluid)
*********"
%exch = mH*(CpH*(THout-THin)+(dps/denH)-to*CpH*log(THout/THin)) %exergy
change in kW"
%*********** Exergy Change due to Temperature *************
% ******************************** Exergy Destruction
******************* "
%exd = -exch-excc % Exergyt destruction

% %%%%%%% EXERGY DESTRUCTION CHECK USING THE ENTROPY GENERATION

%%%%%%%% "

%********************** Exergy Change due to Temperature *************

e1=log(1+Effec*((THin/TCin)-1))
e2=(CpH/Cpc)*log(1-((Cpc*Effec)/(Cpc))*(1-(TCin/THin)))
e=e1+e2
exchtemp=(to*Cmin)*e
exchpress = Pwt+Pwb+Pws
hoursop = 13.7; %Hours of operation in a day"
PPy = Pwtot* hoursop * 365; % Pumping power per year taking only the
tube and shell pressure drops
PPy1=Pwtot1*hoursop*365; % Pumping power per year taking tube, bend
and shell side pressure drops
%Cfuel = 0.12; %Cost of fule kw.h"
Cfuel = 0.05; % Cost of fuel kWh
CPPy= PPy * Cfuel; %Pumping cost per year"
cppy1= PPy1* Cfuel;
%CPPye= PPye * Cfuel;" "Pumping cost per year considering economy"
%interest = 0.10; % Interest of inflation rate for the fuel cost"
interest = 0.18; % Interest of inflation rate for the fuel cost"
multf=1/(interest+1)
%life = 10; % Life of the Heat Exchanger in years"
life = 20; % Life of the Heat Exchanger in years"
% ***************** Cost Associated with Exergy Destruction
************* "
exd = exchtemp + exchpress
Cexdytemp=exchtemp*365*Cfuel;
Cexdypress=exchpress*365*Cfuel;
Cexdy = exd*hoursop*365*Cfuel; %Cost of exergy destruction per year"
% ************ Operating Cost for the life of Heat Exchanger
************** "
opc=Cexdy; % Operating cost including exergy destruction cost"
% ***************TOTAL COST ******************************* "

a11= 10000;
a12= 324;
a13= 0.91; %Values of theses variables a11, a12 and a13 are for
steel-steel Heat Exchanger"

capcost= a11+a12* (af1)^a13 %Capcost considering fouling"

164
Fcr= (interest*((1+interest)^life))/(((1+interest)^life)-1); %Capital
recovery factor
ACC=Fcr*capcost % Annual Capital Cost
Costy=Fcr*capcost+opc; % Total cost per year taking exergy destruction
per year and the annual capital cost
%tcost = capcost+ opc; % Total cost of Heat Exchanger over the life
time of HX considering fouling"
end

% ****************Calculate Fitness****************
if rersViolation==0

% Penalty for negative result

if real(Costy)<0
pNeg=1e7;
else
pNeg=0;
end

%Penalty for complex result

if abs(imag(Costy))>1e-10
pCpx=1e7;
else
pCpx=0;
end

if (pNeg + pCpx )==0

fitness = abs(Costy);
else
fitness = pNeg + pCpx;
end

elseif rersViolation==1
fitness = 5e7;
end

if display==1
disp(sprintf('Length=%2.3f[m]',real(l)))
disp(sprintf('Pitch=%1.3f[m]',real(pt)))
disp(sprintf('Clearance=%0.5f[m]',real(c)))
disp(sprintf('# of Tubes=%4.1f[tubes]',nt))
disp(sprintf('Tube side Velocity=%1.4f[m/s]',velc))
disp(sprintf('Ret=%6.2f[]',ret))
disp(sprintf('Prt=%2.3f[]',prtc))
disp(sprintf('Tube Heat transfer coeff=%4.2f[W/m^2-K]',ht))
disp(sprintf('fc=%1.3f[]',fc))
disp(sprintf('DeltaPt=%3.2f[kPa]',dpt))
disp(sprintf('DeltaPt=%3.2f[kPa]',dpb))
disp(sprintf('as=%1.5f[m^2]',as))
disp(sprintf('De=%1.3f[m]',dse))
disp(sprintf('Shell side Velocity=%1.3f[m/s]',velh))
disp(sprintf('Res=%6.1f[]',res))
disp(sprintf('Prs=%2.3f[]',prth))

165
 disp(sprintf('Shell heat transfer coeff=%4.2f[W/m^2-K]',ho))
disp(sprintf('fs=%1.3f[]',fs))
disp(sprintf('DeltaPs=%3.3f[kPa]',dps))
disp(sprintf('UC=%4.2f[W/m^2-K]',uc))
disp(sprintf('Uf=%4.2f[W/m^2-K]',uf))
disp(sprintf('Uf1=%4.2f[W/m^2-K]',uf1))
disp(sprintf('Af=%4.2f[m^2]',af))
disp(sprintf('Af1=%4.2f[m^2]',af1))
disp(sprintf('Capital Cost=%5.0f[$]',capcost))
disp(sprintf('Total cost Costy=%7.0f[$]',Costy))
disp(sprintf('Operating cost per year=%7.2f[$]',opc))
disp(sprintf('Exergy destruction= %7.3f[kW]',exd))
disp(sprintf('Exergy change in temp parts= %5.3f',exchtemp))
disp(sprintf('Exergy change due to pressure drop=
%5.3f',exchpress))
%disp(sprintf('Exergy change in temp= %5.3f',exchtemp))
disp(sprintf('Cost of exergy destruction per year= %5.3f',Cexdy))
disp(sprintf('Cold fluid outside Temp= %5.3f[K]',TCout))
disp(sprintf('Cmin= %5.3f',Cmin))
disp(sprintf('ACC= %5.3f[$]',ACC))
disp(sprintf('dpt= %5.3f[kPa]',dpt))
disp(sprintf('dpb= %5.3f[kPa]',dpb))
disp(sprintf('dps= %5.3f[kPa]',dps))
disp(sprintf('ho= %5.3f[kw/m k]',ho))
disp(sprintf('ht= %5.3f[kW/m K]',ht))
disp(sprintf('F= %5.3f[]',F))

166
Plate-Fin Heat exchanger
%This program is use to desing a Plate-Fin Heat Exchanger for Energy
Recovery Systems specially in A/C applications"
low rate of the fresh air in cubic meter per second for the installed
tonnage of A/C"
VC=tonnage*0.03159; %[m^3/s]; "Volume flow rate of the fresh air
in cubic meter per second for the installed tonnage of A/C"

relhum_hot=relhum_hotES

%"It does not consider the cost reduction in the manufacturing of A/C
plant"
%"%%%%%%%%%%%%%%%%%%%%%%% SIZIING PROBLEM
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
%"%%%%%%% Hot and Cold fluid properties and conditions
%%%%%%%%%%%%%%"
function fitness=HXG(ind)
global data display trialIX
deltaPHES=ind(1)
deltaPCES=ind(2)
relhum_hotES=ind(3)
lHtES=ind(4)
lctES=ind(5)
rersViolation=0;

tonnage=100;

for HXG_index=1:1,

VH=tonnage*0.03159; %[m^3/s]; %Volume f

THin=313.2%"Hot fluid inlet temperature in Kelvin which is also
fresh air"
TCin=298.2; %"Cold fluid inlet temperature in Kelvin for return
air"
PH=101; %"Inlet pressure on the hot side in kPa"
PC=101; %"Inlet pressure on the cold side in kPa"
deltaPH=deltaPHES
deltaPC=deltaPCES
effec=0.8381;

CpC1 function
[Cmin1, CpC1, CpH1, mH, mC]=Cmint(VH,VC,PH,PC,THin,TCin,relhum_hot)
THout1=THin- effec*(Cmin1/(CpH1*mH))*(THin- TCin)
TCout1=TCin+ effec*(Cmin1/(CpC1*mC))*(THin- TCin);
THavg=(THin+THout1)/2
TCavg=(TCin+TCout1)/2

[Cmin, CpC, CpH]= Cminf(THavg,TCavg,VH,VC,relhum_hot)

TCout=TCin+ effec*(Cmin/(CpC*mC))*(THin- TCin);
THout=303.2
dvisC=(2*10^(-8))*(TCavg)+0.00004
dvisH=(3.923*10^(-8))*THavg+4.461*10^(-6)
kH=(8*10^(-5))*THavg+0.001*(relhum_hot)^2-0.004*(relhum_hot)+0.0036
% Hot fluid thermal conductivity
kC=(4*10^(-5))*TCavg+0.014 % Cold fluid thermal conductivity

167
 %kC=7*(10^-5)*(TCavg)+0.0034

prH=(dvisH*CpH*1000)/kH
prC=(dvisC*CpC*1000)/kC

Mean Density Function

[denHi, denHo, denHm, denCi, denCo, denCm]= denmean(THin, TCin,
THout, TCout,relhum_hot)

CH=(mH*CpH) %"Value of C for Hot fluid at the mean temperature"

CC=(mC*CpC) %"Value of C for cold fluid at the mean temperature"

Cstar=Cmin/CC
ntu=ntu1(effec,Cstar) %" To call the ntu function to calculate the
value of NTU for he given Effcectiveness for a Cross Flow HX with both
fluids unmixed "

ntuC=2*ntu %"This is the initial assumption for initial guess"

ntuH=ntuC %"For a gas-to-gas Heat Exchager it good assumption that
hot and cold fluid NTUs are same"

colfannH= 0.25; %"Ratio of the Colburn and Fanning friction

factor for hot fluid, and initial approximation, colfannH=jH/fH"
colfannC = 0.25; %"Ratio of the Colburn and Fanning friction
factor for cold fluid, an initial approximation, colfannC=jC/fC"

eta0= 0.8; %"Fin efficiency initial guess, becuase we do not

know at that point the value of fin efficiency"

gcc=1;
GH=
((2*gcc*eta0*deltaPH*1000*denHm)/((prH^(2/3))*ntuH)*(colfannH))^0.5;
%"Mass velocity of Hot side fluid kg/m^2.s"

GC=
((2*gcc*eta0*deltaPC*1000*denCm)/((prC^(2/3))*ntuC)*(colfannC))^0.5;
%"Mass velocity of Cold side fluid kg/m^2.s"

geom='H'; %"Selection of a particular geometry, These Geometires

are taken from Kays and London for Offset Fins, ranging from A to U,
Fig-10.52 to Fig-10.71"

Geometric Parameters Function

[pf,b,ls,dh,delta,beta,fta,s,h]= geompara(geom);
%call geompara(geom$:pf,b,ls,dh,delta,beta,fta,s,h)

ReH=(GH*dh)/dvisH %" Hot fluid Reynolds number"

ReC=(GC*dh)/dvisC %" Hot fluid Reynolds number"

[jH, jC, fH, fC]= colbfric (ls,delta,s,h,ReH,ReC);

168
Colburn Factor Function
[jH, jC, fH, fC]= colbfric (ls,delta,s,h,ReH,ReC);
%call colbfric (ls,delta,s,h,ReH,ReC:jH,jC,fH,fC)

%" Calculation of Heat Transfer Coefficients of Hot and Cold

Fluids, Fin Efficiencies and overall Heat Transfer Coefficient "

hHot = (jH*GH*CpH*1000)/((prH)^(2/3));
hCold= (jC*GC*CpC*1000)/((prC)^(2/3));

kf=18; %"Fin Conductivity of Aluminum iconel-625 18 w/m.K"

mfH=((2*hHot/(kf*delta))*(1+(delta/ls)))^(1/2); %"m for the fin

efficiency for the Hot fluid"
mfC=((2*hCold/(kf*delta))*(1+(delta/ls)))^(1/2); %"m for the fin
efficiency for the cold fluid"

lH=(b/2-delta); %"Effective lenght of the fin on the hot side"

lC=(b/2-delta); %"Effective lenght of the fin on the cold side"

etafH=tanh(mfH*lH)/(mfH*lH); %"Hot side fine efficiency"

etafC=tanh(mfC*lC)/(mfC*lC); %"Cold side fine efficiency"

etaoH=(1-(1-etafH)*fta); %"The overall efficiency of fin on the Hot

side"
etaoC=(1-(1-etafC)*fta); %"The overall efficiency of fin on the
Cold side"

deltaw=0.0005; %"Plate thickness"

%" To calculate U (the overall heat transfer coefficient), first

calculate alpha(Hot), alpha(Cold) "

alphaH=(b*beta)/(b+b+2*deltaw); %"Alpha on the hot side"

alphaC=(b*beta)/(b+b+2*deltaw); %"Alpha on the hot side"

alpharatio=(alphaH/alphaC); %"Ratio of the areas and alphas"

%" Now to calcualte the overall heat transfer coefficient "

UH=((1/(etaoH*hHot))+(alpharatio)/(etaoC*hCold))^(-1) %"Overall
heat tranfer coefficient on the hot fluid"

AH=(ntu*Cmin*1000)/(UH)
AC=AH

AoH=(mH/GH) %"From the specified mass flow rate and computed GH,
the minimum free-flow area on the Hot fluid side"
%lHt=(dh*AH)/(4*AoH)
lHt=lHtES

AoC=(mC/GC); %"From the specified mass flow rate and computed GC,
the minimum free-flow area on the Cold fluid side"

169
 lct=lctES
%" To calculate the core frontal area on each fluid side, we first
need to determin the value of sigmal"

sigmaH=((alphaH*dh)/4); %" Hot side fluid sigma which is x-

sectional area to frontal area"
sigmaC=((alphaC*dh)/4); %" Cold side fluid sigma which is x-
sectional area to frontal area"

AfrH=(AoH/sigmaH); %"Frontal area on the hot fluid side"

AfrC=(AoC/sigmaC); %"Frontal area on the cold fluid side"

l3a=AfrC/lHt %"Length form the hot-side fluid frontal area"

l3b=AfrH/lct %"Length form the cold-side fluid frontal area"
l3=l3a
%l3=l3ES
%" At this point we have calculated hot-fluid length (lHt), cold-
fluid length (lCt) and the third dimension (L3)"

%" %%%%%%%%%%%%%% Pressure Drop Calculations

%%%%%%%%%%%%%%%%%%%%%"

ReP=15000; %"Offside strip fin are used on both sides. In such

geometries, because of the frequent boundary layer interruptions, the
flow is well mixed and is treated
% as having the Reynolds number very
Larg approaches to Infinity"
Fintype='square';

KcH, KcC,KeC function ( pressure drop calculation)

[KcH,KcC,KeH,KeC] = KcKe(ReP,sigmaH,sigmaC,Fintype);
%" %%% Correction of (f) factor for the temperature-dependent
property effects, we will first calculate Tw, the thermal resistances
on the hot and cold fluid sides%%%%%"

RH=(1/(etaoH*hHot*AH)); %"Hot side thermal resistance"

RC=(1/(etaoC*hCold*AC)); %"Cold side thermal resistance"

Rratio=(RH/RC); %" Resistance ratio"

Tw=(THavg+Rratio*TCavg)/(1+Rratio);

ThSide='Hot'; %"Thermal side means that is hot fluid (being

cooling), or Cold side (being heating), Value can be put Hot, and Cold"

[mfacH, mfacC]= mfac(ReH,ThSide); %"Call the procedure to have

mfactH, and mfactC"

fHf=(fH*(Tw/THavg)^mfacH); %"Hot side corrected friction factor"

fCf=(fC*(Tw/TCavg)^mfacC); %"Cold side corrected friction factor"

deltaPHc=((GH^2)/(2*denHi))*((1-(sigmaH^2)+KcH)+2*((denHi/denHo)-
1)+(fHf*lHt*denHi)/((dh/4)*denHm)-(1-(sigmaH^2)-
KeH)*(denHi/denHo))*0.001 %"Calculated pressure drop on hot side"

170
 deltaPCc=((GC^2)/(2*denCi))*((1-(sigmaC^2)+KcC)+2*((denCi/denCo)-
1)+(fCf*lct*denCi)/((dh/4)*denCm)-(1-(sigmaC^2)-
KeC)*(denCi/denCo))*0.001 %"Calculated pressure dopr on cold side"

%result$=decide$(deltaPCc,deltaPC)

%" %%%%%%% Sizing Problem completed here

%%%%%%%%%%%%%%%%%%%%%%%%%%% "
%" %%%%%%%% Start of Rating Problem to check this design %%%%%%%"

%"***** Rating problem is to check whether the design in the

Design Problem fulfill the required temperatures or not
****************** "
np=floor(((l3-b-2*deltaw)/(b+b+2*deltaw)))+1; %"We assume there
are np passes for the hot fluid and (np+1) passes for the cold air
having same geometry on both side"
npH=np;
npC=np+1;
%"np=lct/h";
AfrHr=lHt*l3; %"Frontal area on the hot fluid side"
AfrCr=lct*l3; %"Frontal area on the Cold fluid side"

VpHr=lct*lHt*b*np; %"The heat exchager volume between plates on hot

side"
VpCr=lct*lHt*b*(np+1); %"The heat exchager volume between plates on
cold side"

AHr=beta*VpHr; %"The heat transfer area on hot side"

ACr=beta*VpCr; %"The heat transfer area on cold side"

AoHr=(dh*AHr)/(4*lHt); %"Minimum free flow are on hot side"

AoCr=(dh*ACr)/(4*lct); %"Minimum free flow are on hot side"

sigmaHr=(AoHr)/(AfrHr); %"Ratio of minimum free-flow are to

frontal area on hot side"
sigmaCr=(AoCr)/(AfrCr); %"Ratio of minimum free-flow are to frontal
area on cold side"
effecr=0.83; %"Assumed effectiveness for the rating problem"

THoutr=THin-effecr*(THin-TCin); %"Hot fluid outlet temperature"

TCoutr=TCin+effecr*(mH/mC)*(THin-TCin); %"Cold fluid outlet
temperature"
THmr=(THin+THoutr)/2; %"Mean temperature for the hot fluid "
TCmr=(TCin+TCoutr)/2; %"Mean temperature for the cold fluid "

dvisCr=-0.0025*(TCmr)+1.9147;
dvisHr=(3.923*exp(-8))*THmr+4.461*exp(-6);
kHr=(8*exp(-5))*THmr+0.001*(relhum_hot)^2-0.004*(relhum_hot)+0.0036
% Hot fluid thermal conductivity

kCr=(4*exp(-5))*TCmr+0.014; % Cold fluid thermal conductivity

CpCr=0.0006*TCmr+0.835;
CpHr=(0.0043*(relhum_hot)-0.0001)*THmr+(-
1.2097*(relhum_hot)+1.0238)

171
 prHr=(dvisHr*CpHr*1000)/kHr; %"Hot fluid pandtle number"
prCr=(dvisCr*CpCr*1000)/kCr; %"Cold fluid pandtle number"

GHr=(mH/AoHr); %"Hot fluid mass velocity"

GCr=(mC/AoCr); %"Cold fluid mass velocity"

ReHr=(GHr*dh)/dvisHr ; %"Reynolds number on the hot side fluid"

ReCr=(GCr*dh)/dvisCr ; %"Reynolds number on the cold side fluid"

Colburn Factor Function for Rating

[jHr, jCr, fHr, fCr]= colbfricr(ls,delta,s,h,ReHr,ReCr); %"To call
the procedure to have jCr, jHr"

hHotr = (jHr*GHr*CpHr*1000)/((prHr)^(2/3));
hColdr = (jCr*GCr*CpCr*1000)/((prCr)^(2/3));

mfHr=((2*hHotr/(kf*delta))*(1+(delta/ls)))^(1/2); %"m for the fin

efficiency for the Hot fluid"
mfCr=((2*hColdr/(kf*delta))*(1+(delta/ls)))^(1/2); %"m for the fin
efficiency for the cold fluid"

etafHr=tanh(mfHr*lH)/(mfHr*lH); %"Hot side fine efficiency"

etafCr=tanh(mfCr*lC)/(mfCr*lC); %"Cold side fine efficiency"

etaoHr=(1-(1-etafHr)*fta); %"The overall efficiency of fin on the

Hot side"
etaoCr=(1-(1-etafCr)*fta); %"The overall efficiency of fin on the
Cold side"

Aw=lHt*lct*(2*np+2); %"The wall conductance area"

Rw=(deltaw)/(kf*Aw); %"Wall resistance, here kw=kf"

UAr=((1/(etaoHr*hHotr*AHr))+(1/(etaoCr*hColdr*ACr))+Rw)^(-1) ;
%"Overall heat tranfer coefficient on the hot fluid"

ntur=UAr/(Cmin*1000) %"NTU calculation based on rating problem"

effecrc=1-exp((1/Cstar)*((ntur)^0.22)*(exp(-Cstar*((ntur)^0.78))-
1)) %"Effectiveness calculation based on rating problem"

%"NOTE: Now we have to take the logitudinal conduction in the wall.

We evaluae the dercrease in effectiveness dure to longitudinal heat
conduction."

AkH=2*np*lct*deltaw; %"Hot side conduction X-sectional area

for logitudinal heat conduction in the wall"
AkC=2*(np+1)*lHt*deltaw; %"Hot side conduction X-sectional area
for logitudinal heat conduction in the wall"

lamdaH=(kf*AkH)/(lHt*CpH*1000*mH); %"Logitudinal conduction

parameters on the hot fluid"

172
 lamdaC=(kf*AkC)/(lct*CpC*1000*mC); %"Logitudinal conduction
parameters on the cold fluid"

etaratio=(etaoHr*hHotr*AHr)/(etaoCr*hColdr*ACr); %"Ratio of hot and

cold side resistance ratios"
lamdaratio=lamdaH/lamdaC; %"Ratio of Logitudinal conduction
parameter for Hot and Cold Fluid"

effecrcf=effecrc-(0.002)*effecrc; %"Reduction in effectivness due

to Longitudinal Conduction"

Q=effecrcf*(THin-TCin)*Cmin %"Heat Transfer rate in KW"

THoutrf=THin-(Q/(CpH*mH)) %"Hot fluid outlet temperature"

TCoutrf=TCin+(Q/(CpC*mC)) %"Cold fluid outlet temperature"

diffTH= THout-THoutrf %Difference of temperature compared with the

design value"
diffTC=TCout-TCoutrf %Difference of temperature compared with
the design value"

effectf=(THin-THoutrf)/(THin-TCin)

%%%%%%%%%%%%%%%% EXERGY CALCULATIONS %%%%%%%%%%%%%%%%%%%% "

to = 313.2; %Environmental temperature "
% ******* Exergy Change in the Cold Fluid (Return Air) **************"

hcout=0.0646*(TCoutrf^2)-36.098*TCoutrf+5073.6
hcin=0.0646*(TCin^2)-36.098*TCin+5073.6

hci=50.54 % Fixed values used for enthalpies from EES

hco=99.48 % Fixed values used for enthalpies from EES

scout=0.0002*(TCoutrf^2)-0.1126*TCoutrf+21.346
scin=0.0002*(TCin^2)-0.1126*TCin+21.346
%scin=0.0002*(TCin^2)-0.1113*TCin+21.124;
sci=5.789 % Fixed values used for entropy from EES
sco=5.954 % Fixed values used for entropy from EES
excc=mC*((-hco+hci)-to*(-sco+sci))
% ******** Exergy Change in the HotFluid (Fresh Air) **************"

hhout=(0.0412*(relhum_hot)^2+0.1201*(relhum_hot)+0.0009)*THoutrf^2-
(24.634*(relhum_hot)^2+69.069*(relhum_hot)-
0.4985)*THoutrf+(3693.7*(relhum_hot)^2+9953.9*(relhum_hot)-197.47)
hhin=(0.0412*(relhum_hot)^2+0.1201*(relhum_hot)+0.0009)*THin^2-
(24.634*(relhum_hot)^2+69.069*(relhum_hot)-
0.4985)*THin+(3693.7*(relhum_hot)^2+9953.9*(relhum_hot)-197.47)

hhi=146.7
hho=68.03

shout=(0.019*(relhum_hot)+0.0031)*THoutrf+(-
5.5327*(relhum_hot)+4.7715)

173
 shin=(0.019*(relhum_hot)+0.0031)*THin+(-
5.5327*(relhum_hot)+4.7715)
shi=6.108
sho=5.849

exch=mH*((-hho+hhi)-to*(-sho+shi))
% *** Exergy Destruction in Cross Flow Heat Exchanger without Energy
Recovery "
exd= exch+excc
%%% EXERGY DESTRUCTION CHECK USING THE ENTROPY GENERATION %% "

% Entropy Generation in Hot Fluid"

SHgen=mH*((hhout-hhin)/to-(shout-shin)); % Entropy
generation for the hot fluid based on air-vapour mixture with known
humidity ratio"

% Entropy Generation in Cold Fluid"

SCgen=mC*((hcout-hcin)/to-(scout-scin)); % Entropy
generation for the cold fluid based on air-vapour mixture with known
humidity ratio"
%Net Entropy Generation in the System"

Snet = SHgen+SCgen;
%Snet = -(SHgen+SCgen); % Net Entropy Generation in the
System"
% Exergy Destruction"

exdentropy=to*Snet;
%exdentropy=to*Snet ; % Exergy Destruction based on
Entropy Generation"
error= exd-exdentropy;

%%%%%%%%%%%%%%%%%%%%%%% ENERGY RECOVERY FROM ENERYGY RECOVERY

SYSTEM %%%%%%%%%%%%%%%% "

%Qr1 = mH*CpH*(THin-THoutrf) ; % Energy recovered from

or fresh air is cooled by this amount of energy in kW"
Qr1=mH*(hhin-hhout);
COP = 2.53 ; % COP is the coefficient of performace of an
centrally A/C system which is mostly taken as 2.53"

Wc = Qr1/COP; % Compressor work saved which is the exergy saving in

kW"

% COST OF CROSS FLOW PALTE FIN HEAT EXCHANGER FOR ENERGY RECOVERY
SYSTEM *

% %%%%%%% Hot Side Fin Material %%%%%%%% "

denfm = 2770; %Density of fin material in kg/m3 which is
aluminum alloy 2024-T6 with Cu=4.5%, Mg=1.5%, Mn=0.6%"
cfm = 3.25; %Cost of fin material in $/kg"

FmH=(h+pf)*(1/pf)*l3*lct*(floor(lHt/(2*(b+deltaw)))+1)*delta*denfm ;
%Hot side fin material used taking rectangular fin in kg"

174
 CFmH = FmH*cfm; %Cost of Hot side fin material used taking
rectangular fin in $"

%%%%%%% Cold Side Fin Material %%%%%%%% "

FmC = (h+pf)*(1/pf)*l3*lct*(floor((lHt/(2*(b+deltaw))))+2)*delta*denfm
; %Cold side fin material used taking rectangular fin in kg"
CFmC = FmC*cfm; %Cost of Cold side fin material used taking
rectangular fin in $"

% ***************** Cost of Partening Plates ********************** "

denpm = 2770;
%Density of fin material in kg/m3 which is aluminum alloy 2024-t6
with Cu=4.5%, Mg=1.5%, Mn=0.6%"

Ppm=(np+np+2)*lHt*lct*deltaw*denpm;
CPpm = Ppm*cfm ; %Partening plates material cost in $"
Accematcost = 200; %Accessories material cost in $"
Matcost = CFmH + CFmC + CPpm + Accematcost; %
Material cost of heat exchanger"

% ******* Manufacturing Cost ********************** "

n = 3; %Manufacturing cost factor which is usually

determine by the industry most taken as some factor of material cost"
Manfcost = n * Matcost ; %Manufacturing cost of heat echanger"

capcost = Matcost + Manfcost % Total capital cost of energy recovery

system"

% ****** CAPITAL COST BASED ON ESDU-97006 COSTING METHOD

**************** "
Q= mC*CpC*(TCout-TCin); % Heat capacity of the heat
exchanger"
dt1 = THin-TCout;
dt2=THout-TCin;
LMTD = (dt1-dt2)/log(dt1/dt2) ; %Log mean temperature
difference"
Qlmtd = Q*1000/LMTD ; %Heat capacity per unit LMTD
temperatue difference"

[c1,c2,Qlmtd1,Qlmtd2,Bvalue] = Cvalue(Qlmtd,PH)

%call Cvalue(Qlmtd,PH:c1,c2,Qlmtd1,Qlmtd2) " Calling

the Cvalue of program to get C1, C2, Qlmtd1,Qlmtd2"
if (Qlmtd1~=Qlmtd2) %~= This symbol is for not equal to
C
=exp(log(c1)+((log(c1/c2)*(log(Qlmtd/Qlmtd1))/(log(Qlmtd1/Qlmtd2)))));
else
C=exp(log(c1))
end
capcost1= Qlmtd*C
% ********** CAPITAL COST BASED ON Vatavuk COSTING METHOD ********** "
capcost2= 231*(AH+AC)^0.639 % I am using this method in my calculation
because it is based on surface area of heat exchanger

175
% ****** CAPITAL COST BASED ON Ratio Proportion COSTING METHOD **** "
v1 = 1.15*74840/(Bvalue*1000); % Volume based on reference case
v2 = 1.15*Qlmtd/(Bvalue*1000); % Volume based on actual case
capcost3=47000*(v2/v1)^0.6 % Capital cost based on ratio proportion
method
% *********** OPERATING COST ************************** "
% *********** Return air pumping power required ****************** "
effp = 0.8;
% Ppr = ((mC*deltaPCc)/(denCm*effp)) ; %Pumping power required for
cold side (return air) in kW"
Ppr = ((mC*deltaPC)/(denCm*effp));

%**************** Fresh air pumping power required ****************"

% Ppf =((mH*deltaPHc)/(denHm*effp)) ; %Pumping
power required for Hot side (fresh air) in kW"
Ppf =((mH*deltaPH)/(denHm*effp));
%******************* Total Pumping Power Required ****************"
Ppt = Ppr + Ppf ; %Total pumping power required for hot (fresh) and
cold (return) fluid in kW"
% ******** Total pumping cost in US $ ************************** "
hoursop = 18; %Hours of operation in a day"
daysop = 180; % Days of operation in a year"
% Total number of hours of operation is
PPy = Ppt* hoursop * daysop % Pumping power in kWh"
Cfuel = 0.18; % Cost of fuel Euros per kwh"
CPPy = PPy * Cfuel %Pumping cost per year in US $"
interest = 0.10; % Interest of inflation rate for the fuel cost"
multf= 1/(interest+1);
life = 10; % Life of the Heat Exchanger in years"
TPPC = (CPPy * ((multf^life)-1))/(multf-1) ;
% Total pumping cost over the life of the HX in US $ using G.P. seq.
sum** "
%****** Cost Associated with Exergy Destruction in US $
**************** "
exdd = exd * hoursop; %Exergy desturction per day
exdy = exd * hoursop * daysop; %Exergy destruction per
year
Cexdy = exdy * Cfuel; %Cost of exergy destruction per
year"
TexdC = (Cexdy * ((multf^life)-1))/(multf-1);
% Total exergy dest. cost over the life of the HX in US $ using G.P.
seq. sum "
%********* Cost saved in US $ related to Energy Recovery System
********* "
Savenergy = Wc * hoursop * daysop; %Energy saved in a year in kWh"
CSavenergy = Wc * hoursop * daysop* Cfuel;
% Cost of energy saved in a year due to energy recovery system in US $"
Tsavenergy = (CSavenergy * ((multf^life)-1))/(multf-1) %Total cost of
energy saved over the life of the HX in US $ using G.P. seq. sm"
%** Total Operating Cost for the life of Heat Exchager used for
Energy Recovery System ************* "
TOPC = TPPC + TexdC; % Total operating cost including
pumping and exergy destruction cost over the life of heat exchanger"
TOPC1 = TPPC % Total operating cost including only pumping cost
NOPCex = TOPC - Tsavenergy ; % Net operating cost of heat exchanger in
US $"

176
 NOPCp = TOPC1 - Tsavenergy; % Net amount of energy saved due to only
pumping cost
% *** TOTAL COST OF HEAT EXCHANGER FOR ENERGY RECOVERY SYSTEM
******** "
Tcost = TOPC + capcost; %Total cost of heat exchanger
used for engergy recovery system for a Central Air-conditioning system"
% Total cost of heat exchanger including capital and pumping cost.
Tcostwe1P = capcost +TPPC %Total life cycle cost of heat exchanger
including capital and total pumping cost in US $"
Tcostwe2P = capcost1+TPPC %Total life cycle cost of heat exchanger
including capital (ESDU method) and total pumping cost in US $"
Tcostwe3P = capcost2+TPPC %Total life cycle cost of heat exchanger
including capital(Vatavuk) and total pumping cost in US $"
Tcostwe4P = capcost3+TPPC %Total life cycle cost of heat exchanger
including capital (ratio Proportion) and total pumping cost in US $"
Tcostwe1ex = capcost + TOPC %Total life cycle cost of heat exchanger
including capital,total pumping and exergy cost in US $"
Tcostwe2ex = capcost1 + TOPC %Total life cycle cost of heat exchanger
including capital (ESDU method), total pumping and exergy cost in US $"
Tcostwe3ex = capcost2 + TOPC %Total life cycle cost of heat exchanger
including capital, (Vatavuk) total pumping and exergy cost in US $"
Tcostwe4ex = capcost3 + TOPC %Total life cycle cost of heat exchanger
including capital, (ratio Proportion) total pumping and exergy cost in
US $"
% Net amount of energy saved using energy recovery system
% both pumping and exergy destruction cost
Netsaved_energyPump = Tsavenergy - Tcostwe3P
Netsaved_energyPump1=1/Netsaved_energyPump
Netsaved_energyexer= Tsavenergy-Tcostwe3ex
Netsaved_energyexer1=1/Netsaved_energyexer
Pbackperiody= pbp(CPPy,interest,capcost2,CSavenergy)/12
Pbackperiodm= pbp(CPPy,interest,capcost2,CSavenergy)
% ****************Calculate Fitness****************
if rersViolation==0
% % % Penalty for negative result
if real(Netsaved_energyexer1)<0
pNeg=1e7;
else
pNeg=0;
end
% %
% % % Penalty for complex result
if abs(imag(Netsaved_energyexer1))>1e-10
pCpx=1e7;
else
pCpx=0;
end
% %
if (pNeg + pCpx)==0
fitness = abs(Netsaved_energyexer1);
else
fitness = pNeg + pCpx;
end
%
%
elseif rersViolation==1
fitness = 5e7;

177
 end
%
%
if display==1
disp(sprintf('Capital Cost=%6.2f[$]',real(capcost2)))
disp(sprintf('Pumping power per year=%6.2f[kW]',real(PPy)))
disp(sprintf('Pumping cost per year=%6.2f[$]',real(CPPy)))
disp(sprintf('Total pumping cost=%6.2f[$]',real(TOPC1)))
disp(sprintf('Exergy destruction=%4.2f[kW]',real(exd)))
disp(sprintf('Exergy destruction cost per
year=%6.2f[$]',real(Cexdy)))
disp(sprintf('Total exergy destruction cost=%6.2f[$]',real(TexdC)))
disp(sprintf('Total operating cost=%6.2f[$]',real(TOPC)))
disp(sprintf('Save energy per year=%6.2f[$]',real(CSavenergy)))
disp(sprintf('Total save energy=%6.2f[$]',real(Tsavenergy)))
disp(sprintf('Netsaved_energy taking Pumping cost
only%6.3f[$]',real(Netsaved_energyPump)))
disp(sprintf('Netsaved_energy taking pumping and
exergy=%6.2f[$]',Netsaved_energyexer))
disp(sprintf('l3=%2.2f[m]',real(l3)))
disp(sprintf('lHt=%2.3f[m]',real(lHt)))
disp(sprintf('lct=%2.3f[m]',real(lct)))
disp(sprintf('Paybackperiody=%3.2f[years]',Pbackperiody))
disp(sprintf('Paybackperiodm=%3.2f[months]',Pbackperiodm))
disp(sprintf('AC=%3.2f[m^2]',real(AC)))
disp(sprintf('AH=%3.2f[m^2]',real(AH)))
disp(sprintf('excc=%6.2f[m^2]',real(excc)))
disp(sprintf('exch=%6.2f[m^2]',real(exch)))
disp(sprintf('Tcostwe3ex=%6.2f[m^2]',real(Tcostwe3ex)))
disp(sprintf('THoutrf=%3.2f[$]',real(THoutrf)))
disp(sprintf('TCourtrf=%3.2f[$]',real(TCoutrf)))
end
end

178
 Tube-Fin Heat Exchanger
%This program is use to desing a Tube-Fin Heat Exchanger For Radiator"
% Fluid on Cold side is water"
% Fluid on hot side is gas having properties of air
%"%%%%%%%%%%% SIZIING PROBLEM %%%%%%%%%%%%%%"
%"%%%% Hot and Cold fluid properties and conditions %%%%%%%%%%%%%"

function fitness=HXG(ind)
global data display trialIX
sigmaES=ind(1)
alphaES=ind(2)
dhES=ind(3)
doES=ind(4)
LfES=ind(5)
PfES=ind(6)
tES=ind(7)
AftaES=ind(8)
hES=ind(9)
dES=ind(10)
mjES=ind(11)
mfES=ind(12)
cjES=ind(13)
cfES=ind(14)

rersViolation=0;
for HXG_index=1:1,

mH=1.25 % Hot side mass flow rate in kg/s

mC=1; % Hot side mass flow rate in kg/s
ht=0.335 %Height of exchanger
w=0.6 % Width of exchanger
Afr = ht*w % Frontal are in square meters
kt= 237 % Thermal conductivity of Aluminum in w/m-K
kf=237 % Thermal conductivity of aluminum in w/m-K

THin=825 % Hot fluid inlet temperature in Kelsvin which is exaust gases

(properties similar to air)
TCin=290; % Cold fluid inlet temperature in Kelvin for fresh air
air"
TCout=370 ; % Cold fluid outlet temperature in Kelvin for Water"

[Cmin1, CpC1, CpH1]=Cmint(mH,mC,THin,TCin)

%TCout1=((mH*CpH1)*(THin-THout))/(mC*CpC1)+TCin %Cold fluid outlet
temperature at inlet temperature
THout1=THin-((mC*CpC1)*(TCout-TCin))/(mH*CpH1)
THavg=(THin+THout1)/2
TCavg=(TCin+TCout)/2
[Cmin, Cmax, CpC, CpH]=Cminf(THavg,TCavg,mH,mC)

%* Calculation of Cold Fluid Outlet Temperature at Cp values at Tavg

**"
%TCout=((mH*CpH)*(THin-THout))/(mC*CpC)+TCin %Cold fluid outlet
temperature at average temperature
THout=THin-((mC*CpC)*(TCout-TCin))/(mH*CpH)
% Hot fluid outlet temperature at average temperature

179
 %Calculation of dynamic viscosities of hot (Ehylene Glycole) and
cold fluid (Air)"

dvisC=(4.13949255084809E-15)*(TCavg^6)-(8.34965766217179E-
12)*(TCavg^5)+(7.02375037618649E-09)*(TCavg^4)-(3.15496072315551E-
06)*(TCavg^3)+(0.000798437530456707)*(TCavg^2)-
(0.108000640247946)*(TCavg)+6.10514317010975

dvisH=-1.41045051934921e-16*(THavg^4)+4.26038307798033e-
13*(THavg^3)-4.9185906918802e-10*(THavg^2)+2.86556105638087e-7*(THavg)-
3.84040920614829e-5

kC=(4.65098440380754E-14)*(TCavg^6)-(9.34636114311804E-
11)*(TCavg^5)+(7.84847947075993E-08)*(TCavg^4)-
(0.0000352236122081899)*(TCavg^3)+(0.0088936066460256)*(TCavg^2)-
(1.1933024986979)*(TCavg)+66.6939160860926
kH=-5.8890931849554e-14*(THavg^4)+1.78928079797e-10*(THavg^3)-
2.22228703238359e-7*(THavg^2)+1.85541881867172e-4*(THavg)-
1.65393414701246e-2
prH=(dvisH*CpH*1000)/kH % Hot fluid Prandtl number
prC=(dvisC*CpC*1000)/kC % Cold fluid Prandtl number
denH=2.74123242870646e-12*(THavg^4)-9.319335988338889e-
9*(THavg^3)+1.25346957767038e-5*(THavg^2)-8.32071288733361e-
3*THavg+2.71856873676652

denHi=2.74123242870646e-12*(THin^4)-9.319335988338889e-
9*(THin^3)+1.25346957767038e-5*(THin^2)-8.32071288733361e-
3*THin+2.71856873676652

% Cold fluid density at inlet temperature

denCi=-4.00292558369689E-11*(TCin^6)+7.97535384043824E-
08*(TCin^5)-0.0000661657464616363*(TCin^4)+0.0292672241773225*(TCin^3)-
7.28526352257204*(TCin^2)+968.238056843599*TCin-52693.4274990621

denHo=2.74123242870646e-12*(THout^4)-9.319335988338889e-
9*(THout^3)+1.25346957767038e-5*(THout^2)-8.32071288733361e-
3*THout+2.71856873676652

denCo=-4.00292558369689E-11*(TCout^6)+7.97535384043824E-
08*(TCout^5)-
0.0000661657464616363*(TCout^4)+0.0292672241773225*(TCout^3)-
7.28526352257204*(TCout^2)+968.238056843599*TCout-52693.4274990621
denC=-4.00292558369689E-11*(TCavg^6)+7.97535384043824E-
08*(TCavg^5)-
0.0000661657464616363*(TCavg^4)+0.0292672241773225*(TCavg^3)-
7.28526352257204*(TCavg^2)+968.238056843599*TCavg-52693.4274990621

CH=(mH*CpH) %"Value of C for Hot fluid at the mean temperature"

CC=(mC*CpC) %"Value of C for cold fluid at the mean temperature"
Cstar=Cmin/Cmax;
effec=(CH*(THin-THout))/(Cmin*(THin-TCin))

180
NTU Function
[ntu]= ntu1(CpH,CpC,mH,mC,Cstar,effec) %" To call the ntu function to
calculate the value of NTU for he given Effcectiveness for a Cross Flow
HX with both fluids unmixed "

%geom='A'%"Selection of a particular geometry, These Geometires are

taken from Kays and London for Offset Fins, ranging from A to F
%[sigma alpha dh do di df Pf t Afta h d]= geompara(geom)
sigma=sigmaES
alpha=alphaES
dh=dhES
do=doES
di=doES-0.0025; % Inner tube diameter in meters
df=2*LfES+doES
Pf=PfES
t=tES
Afta=AftaES
h=hES
d=dES
mj=mjES
mf=mfES
cj=cjES
cf=cfES

ntH=floor(ht/h) % No tubes in vertical direction

ReC=(4*mC)/(3.141*di*dvisC*ntH) % Inside fluid the properties of
water
GH=(mH)/(sigma*Afr) %Value of mass valocity for hot fluid in
kg/m^2.s
ReH=(GH*dh)/dvisH %" Hot fluid Reynolds number"
% For hot fluid
jH=cj*(ReH)^(-mj)
fH=cf*(ReH)^(-mf)
% For cold fluid
jC=cj*(ReC)^(-mj)
fC=cf*(ReC)^(-mf)
hHot=(jH*GH*CpH*1000)/((prH)^(2/3)) % outside heat transfer coefficient
for hot gases (properties of Air)

Heat Transfer Coefficient on Cold Side

hCold1= hCold(ReC,kC,di,prC) % Inside or Cold side heat tansfer
coefficient for water
% Inside or Hot side heat transfer coefficient
AH= 1.5707*t*(df-do); % Inside surface area of the fin
Lf=LfES % Fin lenght in m
%Lf= (df-do)/2; %Fin length
P=3.141*(df+do) %Perimeter
M= ((hHot*P)/(AH*kf))^(0.5); % M for the fin efficiency
etao= tanh(M*Lf)/(M*Lf); % Fin efficiency
AhAc=(di)/(0.5*(1/Pf)*(df^2-do^2)+do-(1/Pf)*do*t); % Ratio of hot
and cold sides areas

u=((1/(hCold1*AhAc))+(di*log(do/di))/(2*kt*AhAc)+(1/(hHot*etao)))^(-1);
% Overall heat transfer coefficients

181
 %Vtot=(Ac/alpha); % Total volume of radiator
Vtot1= Vtot(ntu,mH,mC,CpC,CpH,alpha,u)
depth=Vtot1/Afr % Depth of radiator in meters
%NTP = Afr/h % Number of tubes in a pass
NP= ceil(depth/d)+1 % Number of passes in depth
NP1=NP-1
%NT=NTP*NP % Total number of tubes in a radiator
NT=(ntH*(NP-1)+NP1*(ntH-1))
% ************************ Pressure drop on fin side
*****************
V=ht*w*depth;
Af=(df-do)*t;
decfac=(1+sigma^2)*((denHi/denHo)-
1)+(fH*alpha*Vtot1*denHi)/(sigma*Afr*denH)
if (decfac<0) %Selection of friction factor
deltaPH=0
else
deltaPH= ((GH^2/(2*denHi)))*((1+sigma^2)*((denHi/denHo)-
1)+(fH*alpha*Vtot1*denHi)/(sigma*Afr*denH))
end
% **************** Pressure drop on Tube side *****************
%[deltaPH]=deltaPh(ReH,di,mH,denH,w,NP,ntH)

Function for Power Losses Hot and Cold side Delta

% Power losses on hot and cold side
effP=0.80
PH = (mH*deltaPH)/(denH*effP) % Hot side power loss in watts
PC = (mC*deltaPC1)/(denC*effP) % Cold side power loss in watts
PPower=(PH+PC)/1000 % Total pumping power loss in kwatts
%******** Effectiveness of Heat Exchanger ************************ "
%%%%%%%% EXERGY CALCULATIONS %%%%%%%%%%%%%%%%%%%% "
to = 300; %Environmental temperature "
% Exergy Change in the Cold Fluid (Fresh air taken as an Ideal Gas)
****"
PHin=101; % Inlet pressure of air in atmospheric pressure in
kPa
PHout=PHin-(deltaPH/1000) % Outlet pressure of hot air in kPa
exH=mH*(CpH*(THout-THin)-to*((CpH*log(THout/THin)-
(8.314/29)*log(PHin/PHout))))
% Exergy desturction on cold side in kW
% *** Exergy Change in the HotFluid (hot air Ethylene Glycole)
********"
exC=mC*(CpC*(TCout-TCin)+(deltaPC1/(denC*1000))-CpC*to*log(TCout/TCin))
% Exergy destruciton on hot side in kW
exd = -exH-exC % Net Exergy desturction in kW
%*** COST OF CROSS FLOW TUBE FIN HEAT EXCHANGER FOR CAR RADIAT
********** "
% % Note: All Costs are in US $"
%********** CAPITAL COST ************************ "
AC= pi()*(do)*w*NT % Surface area on the cold side (aluminium)
AHf = (pi()*(df^2-do^2)/4+pi()*df*t)*((NT*w)/(Pf))
% Surface area on the hot side (Fin side Aluminum))
capcost1 = 231*(AHf+AC)^0.639
% Using Vatavuk Method for captial cost which is based on
surface area of heat exchanger

182
capcost2 = 231*(AHf+AC)^0.639 % Using Vatavuk Method for captial cost
which is based on surface area of heat exchanger
%
hoursop = 18; %Hours of operation in a day"

daysop = 180; % Days of operation in a year"

% Total number of hours of operation is

PPy = PPower* hoursop * daysop % Pumping power in kWh"

Cfuel = 0.12; % Cost of fuel Euros per kwh"

CPPy = PPy * Cfuel %Pumping cost per year in US $"

interest = 0.10; % Interest of inflation rate for the

fuel cost"

multf= 1/(interest+1);

life = 10; % Life of the Heat Exchanger in years"

TPPC = (CPPy * ((multf^life)-1))/(multf-1) ; % Total

pumping cost over the life of the HX in US $ using G.P. seq. sum** "

%********** Cost Associated with Exergy Destruction in US $

************ "
exdd = exd * hoursop; %Exergy desturction per day
exdy = exd * hoursop * daysop; %Exergy destruction per year
Cexdy = exdy * Cfuel; %Cost of exergy destruction per
year"

TexdC = (Cexdy * ((multf^life)-1))/(multf-1); % Total

exergy dest. cost over the life of the HX in US $ using G.P. seq. sum "

TOPC1=TPPC % Total operating cost including only pumping cost

TOPC2=TPPC+TexdC % Total operating cost including both pumping and
exergy destruction cost.
TcostP=TOPC1+capcost2 % Total lifecycle cost based on pumping and
capital cost
Tcost=TOPC2+capcost2 % Total lifecycle ecost based on pumping and
capital cost
TcostPE=TOPC2+capcost2 % Total lifecycle cost based on puming, exergy
destruction and capital cost.
%Tcost= TOPC+capcost2 % Total cost for the whole lifecycle.
dhf=abs(hHot-hCold1)
% % ****************Calculate Fitness****************
if rersViolation==0
% % %
% % % % Penalty for negative result
if real(Tcost)<0
pNeg=1e7;
else
pNeg=0;

183
 end
% % %
% % % % Penalty for complex result
if abs(imag(Tcost))>1e-10
pCpx=1e7;
else
pCpx=0;
end
if (pNeg + pCpx)==0
fitness = abs(Tcost);
else
fitness = pNeg + pCpx;
end
% %
% %
elseif rersViolation==1
fitness = 5e7;
end

if display==1
disp(sprintf('Capital Cost=%4.2f[$]',real(capcost2)))
disp(sprintf('Pumping Cost per year=%6.2f[$]',real(CPPy)))
%disp(sprintf('Pumping cost per year=%6.2f[$]',real(CPPy)))
disp(sprintf('Total pumping cost=%6.2f[$]',real(TPPC)))
disp(sprintf('Exergy destruction year=%4.2f[kW]',real(exdy)))
disp(sprintf('Exergy destruction cost per
year=%6.2f[$]',real(Cexdy)))
disp(sprintf('Total exergy destruction
cost=%6.2f[$]',real(TexdC)))
disp(sprintf('Total operating cost=%6.2f[$]',real(TOPC2)))
disp(sprintf('Tcost=%5.2f[m^2]',real(Tcost)))
disp(sprintf('AC=%3.2f[m^2]',real(AC)))
disp(sprintf('AH=%3.5f[m^2]',real(AHf)))
disp(sprintf('efffHX=%1.4f[m^2]',real(effec)))
disp(sprintf('ntu=%3.4f[m^2]',real(ntu)))
disp(sprintf('NP=%3f[Passes]',real(NP)))
disp(sprintf('NT=%3f[# of Tubes]',real(NT)))
disp(sprintf('Vtot1=%2.4f[m^3]',real(Vtot1)))
disp(sprintf('dh=%2.4f[m^3]',real(dhf)))

end
end

184
Evolutionary Algorithm

Main Evolutionary Algorithm Function

function [Pop,PopHistory] = myES(opttask,esopt)

% opttask : optimization task

% opttask.func : handle on function to be minimized
% opttask.dim : dimension n of optimization problem
% optask.range : range of objectparameters (n x 2) matrix

% esopt : options of evolution strategie

% esopt.mue : population size
% esopt.lambda : number of offspring
% esopt.maxGenerations : maximum number of function calls
% esopt.recombination : define type of recombination from
('no','intermedian','discrete')
% esopt.rho : number of parents used to generate one offspring
% individual; has to be defined if esopt.recombination
is
% set to 'intermedian' or 'discrete'
% esopt.selection : define selection method as one of
('plus','komma')

% initialize population
Pop.individuals = zeros(opttask.dim,esopt.mue);
Pop.fit = zeros(1,esopt.mue);
Pop.sigma = zeros(opttask.dim,esopt.mue);
for k = 1:esopt.mue
Pop.individuals(:,k) = opttask.range(:,1) + (opttask.range(:,2)-
opttask.range(:,1)).*rand(opttask.dim,1);
Pop.fit(k) = feval(opttask.func,Pop.individuals(:,k));
Pop.sigma(:,k) = 0.001*(opttask.range(:,2)-opttask.range(:,1));%<--
-
end
[Pop.fit IX] = sort(Pop.fit);
Pop.individuals = Pop.individuals(:,IX);
Pop.sigma = Pop.sigma(:,IX);
Pop.generation = 0;
Pop.best = Pop.individuals(:,1);
Pop.bestFit = Pop.fit(1);
% further initializations
par.tao0 = (sqrt(2*opttask.dim))^-1; % for self adaption
par.tao = (sqrt(2*sqrt(opttask.dim)))^-1;

esopt.PreLambda = esopt.lambda;

% save data history

PopHistory{Pop.generation+1} = Pop;

% start es loop
for g = 1:esopt.maxGenerations
Pop.generation = g;
% generate offspring (Recombination)
switch esopt.recombination

185
 case('no')
Pop = Reproduction(Pop,esopt);
case('intermedian')
Pop = RecombinationIntermedian(Pop,esopt);
case('discrete')
Pop = RecombinationDiscrete(Pop,esopt);
otherwise
error('recombination method not definend')
end
% mutate offspring
Pop.PreOffspringSigma = Pop.PreOffspringSigma .*
exp(par.tao0*ones(opttask.dim,1)*randn(1,esopt.PreLambda) +
par.tao*randn(opttask.dim,esopt.PreLambda));
Pop.PreOffspring = Pop.PreOffspring +
Pop.PreOffspringSigma.*randn(opttask.dim,esopt.PreLambda);
Pop.PreOffspring =
max(Pop.PreOffspring,opttask.range(:,1)*ones(1,esopt.PreLambda));
Pop.PreOffspring =
min(Pop.PreOffspring,opttask.range(:,2)*ones(1,esopt.PreLambda));
% real fitness evalation
Pop.offspring = Pop.PreOffspring;
Pop.offspringSigma = Pop.PreOffspringSigma;
Pop = rmfield(Pop,'PreOffspring');
Pop = rmfield(Pop,'PreOffspringSigma');
Pop.offspringFit = zeros(1,esopt.lambda);
for k = 1:esopt.lambda
Pop.offspringFit(k) =
feval(opttask.func,Pop.offspring(:,k));
end
% selection (select mue best individuals from lambda (+ mue)
offspring)
[Pop.offspringFit IX] = sort(Pop.offspringFit); % sorting offspring
according to fitness
Pop.offspring = Pop.offspring(:,IX);
Pop.offspringSigma = Pop.offspringSigma(:,IX); + λ
switch esopt.selection
case 'plus'
Pop = PlusSelection(Pop,esopt);
case 'komma'
Pop = KommaSelection(Pop,esopt);
otherwise
error('selection method not defined')
end
if Pop.fit(1) < Pop.bestFit
Pop.best = Pop.individuals(:,1);
Pop.bestFit = Pop.fit(1);
end
% save data history
PopHistory{Pop.generation+1} = Pop;
end % end of ES- loop

(µ + λ) Selection Function
function Pop = PlusSelection(Pop,esopt)

allIndividuals = [Pop.individuals,Pop.offspring];
allSigmas = [Pop.sigma, Pop.offspringSigma];

186
allFitness = [Pop.fit, Pop.offspringFit];

[allFitness IX] = sort(allFitness);

allIndividuals = allIndividuals(:,IX);
allSigmas = allSigmas(:,IX);

Pop.individuals = allIndividuals(:,1:esopt.mue);
Pop.fit = allFitness(:,1:esopt.mue);
Pop.sigma = allSigmas(:,1:esopt.mue);

Recombination Function Using Intermedian

Function Pop = RecombinationIntermedian(Pop,esopt)

[dim,j] = size(Pop.individuals);
% generate lambda offspring individuals
Pop.PreOffspring = zeros(dim,esopt.PreLambda);
Pop.PreOffspringSigma = zeros(dim,esopt.PreLambda);
for i = 1:esopt.PreLambda
% select rho parents to create offspring individual
ind = ceil(esopt.mue*rand(1,esopt.rho));
% create offspring
Pop.PreOffspring(:,i) = sum(Pop.individuals(:,ind),2)./esopt.rho;
Pop.PreOffspringSigma(:,i) = sum(Pop.sigma(:,ind),2)./esopt.rho;
end

Reproduction Function

function Pop = Reproduction(Pop,esopt)

IX = ceil(esopt.mue*rand(1,esopt.PreLambda));
Pop.PreOffspring = Pop.individuals(:,IX);
Pop.PreOffspringSigma = Pop.sigma(:,IX);

187
 REFERENCES

[1] H. P. Schwefel, Numerical Optimization of Computer Models, Wiley,

Chichestser, 1981.

[2] Anonio C. Caputo, Pacifico M. Pelagagge et al, Heat exchanger design on

economic optimization, Applied Thermal Engineering, 28, pp 1151-1159, 2007.

[3] Yavuz Őzçelik, Exergetic optimization of shell and tube heat exchangers using
genetic based algorithm, 27, pp 1849-1856, 2007

[4] R.K. Shah, K.J Bell, The CRC Handbook of Thermal Engineering, CRC Press,
2000.

[5] Sonntag, Borgnakke Van Wylen, Fundamentals of Thermodynamics, 6th Edition,

John Wiley & Sons, Inc, 2003

[6] Tozer R, Valero A, Lozano, M.A.. Thermo-economics Applied to HVAC

Systems, ASHRAE Transactions. 105, Part 1: pp.1247-1255, 1999

[7] Sommerer F,. Evaluation of Absorption Cycles with Respect to COP and
Economics. International Journal of Refrigeration. 19(1): pp.19-24, 1996.

[8] Tsatsaronis G, Lin L, Tawfik T, Gallaspy D T, Exergoeconomic Evaluation of a

KRW-based IGCC Power Plant. Journal of Engineering for Gas Turbines and
Power. 116(2): pp.300-306, 1994

[9] Zalewski W, Niezgoda Z B, Litwink M., Optimization of Evaporation Fluid

Coolers. International Journal of Refrigeration. 23(7): pp.553-565, 2000

[10] Fowler, A.J Correlation of Optimal Sizes of Bodies with External Forced
Convection Heat Transfer. International Communications in Heat and Mass
Transfer. 21(1): pp.17-27, 1994.

[11] D.Q. Kern, Process Heat Transfer, McGraw-Hill, 1950.

[12] R.K. Sinnot, Coulson and Richardson‟s Chemical Engineering, Chemical

Engineering Design, vol 6, Butterworth-Heinemann, 2005.

[13] Prabhata K. Swamee, Nitin Aggarwal, Optimum design of double pipe heat
exchanger, International Journal of Heat and Mass Transfer, 51, pp 2260-2266,
2008.

[14] Resat Selbas, Onder Kizilkan et al, A new design approach for shell-and-tube heat
exchangers using genetic algorithms from economic point of view, Chemical
Engineering and Processing 45, pp 268-275, 2006

188
[15] Jose M. Ponce-Ortega et al, Use of genetic algorithms for optimal design of shell-
and-tube heat exchangers, Applied Thermal Engineering, 2008

[16] Philippe Wildi-Tremblay and Louis Gosselin, Minimizing shell-and-tube heat

exchanger cost with genetic algorithms and considering maintenance, Int. J.
Energy Res. 31: pp 867-885, 2007

[17] Arzu Sencan, Kemal A. Yakut et.al, Exergy analysis of lithium bromide/water
absorption systems, Renewable Energy, 30: pp 645-657, 2004.

[18] Raj M. Manglik, Arthur E. Bergies, Heat transfer and pressure drop correlations
for the rectangular offset strip fin compact heat exchanger, Experimental thermal
and fluid science, 10, pp 171-180, 1995.

[19] Aytunc Erek, Baris Ozerdem et.al. Effect of geometrical parameters on heat
transfer and pressure drop characteristics of plate fin and tube heat exchangers,
Applied Thermal Engineering, 25, pp. 2421-2431, 2005.

[20] Yonghan Kim, Yongchan Kim, Heat transfer characteristics of flat plate finned-
tube heat exchangers with large fin pitch, International Journal of Refrigeration,
28; pp 851-858, 2005

[21] George Tsatsronis, Moung-Ho Park, On avoidable and unavoidable exergy

destructions and investment costs in thermal systems, Energy Conversion and
Management, 43, pp 1259-1270, 2002

[22] R.L. Cornelissen, G.G. Hirs, Exergetic optimization of a heat exchanger, Energy
Conversion and Management, 38, pp 15-17, 1997

[23] Hao Peng, Xiang Ling, Optimal design approach for the plate-fin heat exchangers
using neural networks cooperated with genetic algorithms, Applied Thermal
Engineering, 28, pp 642-650, 2008.

[24] J.M. Ponce-Ortega, M. Serna-Gonzalez, et.al, Design and optimization of

multipass heat exchangers, 47, pp 906-913, 2008.

[25] C. Arslanturk, A. Feridum Ozguc, Optimization of a central – heating radiator,

Applied Energy , Volume No. 83, pp 1190 – 1197, 2006

[26] Dong-Keun Yang, Kwan-Soo Lee, et.al., Fin spacing optimization of a fin-tube
heat exchanger under frosting conditions, International Journal of Heat and mass
transfer, 49, pp 2619-2625, 2006.

189
[27] Sanjay K. Roy, and Branko L. Avanic, Optimization of a space radiator with
energy storage, International communication in Heat and Mass Transfer, 33, pp.
544-551, 2006.

[28] D. Ganga. Charyulu, G. Singh, J.K. Sharma, Performance evaluation of a radiator

in a diesel engine – a case study, Volume No. 19, pp 625 – 639, 1999.

[29] V. Singh, V. Aute, R. Radermacher, A heat exchanger model for air-to-

refrigerant fin-and-tube heat exchanger with arbitrary fin sheet, International
journal of refrigeration, Volume No. 32, pp 1724 – 1735, 2009

[30] H. T. Chen, W.L. Hsu, Estimation of heat transfer coefficient on the fin of
annular – finned tube heat exchangers in natural convection for various fin
spacing, International journal of heat and mass transfer, Volume No. 50, pp 1750
– 1761, 2007

[31] B. Kundu and P. K. Das, Optimum dimensions of plate fins for fin – tube heat
exchangers, Int. J. Heat and Fluid Flow, Volume No. 18, pp530 – 537,1997

[32] C. Arslanturk, Optimum design of space radiators with temperature-dependent

thermal conductivity, Applied thermal engineering, Volume No. 26, pp 1149-
1157,2006

[33] Z. Gang Qi, Jiang-ping Chen, Zhi-jiu Chen, Parametric study on the performance
of a heat exchanger with corrugated louvered fins, Applied thermal engineering,
Volume No. 27, pp 539 – 544, 2007.

[34] J. H. Holland, Adaptation in Natural and Artificial Systems, MIT Press,

Cambridge, MA, 1992.

[35] D. E. Goldberg, Genetic Algorithm in Search Optimization and Machine

Learning, Addison-Wesely, Reading, MA, 1989.

[36] I. Rechenberg, Evolutionsstrategie; Optimierung Technischer Systeme nach

Prinzipien der Biologischen Evolution, Frommann-Holzboog, Stuttgart, 1973.

[37] L. J. Fogel, A. J. Owens, and M. J. Walsh, Artificial Intelligence through

Simulated Evolution, Wiley, New York, 1966.

[38] D. B. Fogel, Evolutionary Computation: Toward a New Philosophy of Machine

Intelligence, IEEE Press, Piscataway, NJ, 1995.

[39] Hans-Paul Schwefel, Hans-Georg Beyer, Evolution Strategies a comprehensive

introduction, National Computing 1: pp, 3-52, 2002

190
[40] Beyer H-G, The Theory of Evolution Strategies, National Computing Series,
Springer, Heildelberg, 2001.

[41] Rechenberg I, Evolutionsstrategie‟94. Frommann-Holzboog Verilag, Stuttgart,

1994

[42] P.S. Lee, S.V. Garimlla, D. Liu, Investigation of heat transfer in rectangular micro
channels, International Journal of Heat and Mass Transfer 48, 1688-1704, 2005.

[43] G.F. Hewitt (Ed.), Heat Exchanger Design Handbook, Begell House, New York,
1998.

[44] R.K. Sinnott, Coulson and Richardson‟s Chemical Engineering, Chemical

Engineering Design, vol. 6, Butterworth-Heinemann, 2005

[45] M. Taaal, I. Bulatov, J. Klemes, P. Stehlik, Cost estimation and energy price
forecast for economic evaluation of retrofit projects, Applied Engineering, 23,
1819-1835, 2003

[46] W.M. Kays and A.L. London , Compact Heat Exchangers, 3rd Edition, John
Wiley and sons, 1998

[47] Xie GN, Sunden B, Wang Q W. Optimization of Compact Heat Exchangers by a

Genetic Algorithm, Applied Thermal Engineering, 28, 895-906, 2008.

[48] Ramesh K. Shah and Dusan P. Sekulic‟, Fundamentals of Heat Exchanger

Design, New Jersey, John Wiley & Sons, Inc, pp, 393-419, 2003

[49] Meshra, M., Das, P. K., and Sarangi, S., Second Law Based Optimization of
Cross Flow Plate–Fin Heat Exchanger Design Using Genetic Algorithm, Applied
Thermal Engineering, 29, 2983-2989, 2009

[50] Pouria Ahmadi, Hassan Hajabdollahi, Ibrahim Dincer, Cost and Entropy
Generation Minimization of a Cross-Flow Plate-Fin Heat Exchanger using Multi-
Objective Genetic Algorithm, Journal of Heat Transfer, 133, 021801-1-021801-
10, 2011

[51] Carrier Hand book of air-conditioning, 2010.

[52] Incopera, DeWitt et al., Fundamental of heat and mass transfer, 6th Edition, John
Wiley & Sons,2006

[53] Mor, Y,. and Nakayama, W., Recent Advances in Compact Heat Exchanger in
Japan , in Compact Heat Exchangers, History, Technological Advancement and
Mechanical Design Problems, R.K. Shah et al. HTD-Vol 10, pp. 5-16, ASME,
New York, 1980.

191
[54] Lee, K.-B., and Kwon, Y,-K., Flow and Thermal Field with Relevance to Heat
Transfer Enhancement of Interrupted-Plate Heat Exchangers, Exp. Heat Transfer
5, 83-100. 1992.

[55] F.E.M. Saboya, E.M. Sparrow, Transfer Characteristics of two-row plate fin and
tube heat exchanger configurations, Int. J. Heat Mass Transfer 19, pp. 41-49,
1976.

[56] L.A.O. Rocha, F.EM. Saboya, J.V.C. Vargas, A comparative study of elliptical
and circular sections in one and two row tubes and plate fin heat exchangers, Int.
J. Heat Fluid Flow , 18, pp. 247-252, 1997

[57] D.K. Yang, K.S. Lee, Dimensionless correlations of frost properties on a cold
plate, Int. J. Refrigeration 27(1), pp. 89-96, 2004.

[58] D.K. Yang, K.S. Lee, D.J. Cha, Frost Formation on a cold surface under turbulent
flow, Int. J. Refrigeration 29 (2), pp 164-169, 2006.

[59] Brwon, A, Optimum dimensions of uniform annular fins. Int. J. Heat Mass
Transfer, 8, pp. 655-662, 1965.

[60] Zbronsky, H, Temperature distribution and efficiency of a heat exchanger using

square fins on round tubes, J. Appl. Mech., 22, pp. 119-122, 1955

[61] Bilar A. Qureshi, Syed M. Zubari, Second-law-based performance evaluation of

cooling towers and evaporative heat exchangers, International Journal of Thermal
Sciences, 46, pp. 188-198, 2007.

[62] Chang Nyung Kim, Jing Jeong, et.al., Evaluation of thermal contact conductance
using a new experimental-numerical method in fin-tube heat exchangers,
International Journal of Refrigeration, 26, pp. 900-908, 2003.

[63] Dong Junqi, Chen Jiangping, et.al. Heat transfer and pressure drop corrections for
the wavy fin and flat tube heat exchangers, Applied Thermal Engineering, 27, pp.
2066-2073, 2007.

[64] Lieke Wang, Bengt Sunden, Optimal design of plate heat exchangers with and
without pressure drop specifications, Applied Thermal Engineering, 23, pp 295-
311, 2003

[65] S.P. Wang, Q.L. Chen, et.al. A phenomenological equation of exergy transfer and
its application, Energy, 30, pp 85-95, 2005.

[66] Jorge, A. W. Gut, et.al. Optimal configuration design for plate heat exchangers,
International Journal of Heat and Mass Transfer, 47, pp 4833-4848, 2005.

192
[67] Bayram Sahin, Kenan Yakut, et.al, Optimum design parameters of a heat
exchanger, Applied Energy, 82, pp 90-106, 2005

[68] J.M. Reneaume, N. Niclout, MNLP Optimization of Plate fin Heat Exchangers,
Chem. Biochem Eng. Q. 17 (1), pp 65-76, 2003

[69] Y.B. Tao, Y.L. He et.al, Three-dimensional numerical study of wavy fin-and-tube
heat exchangers and field synergy principle analysis, International Journal of Heat
and Mass Transfer, 50, pp 1163-1175, 2007

[70] Worachest Pirompugd, Somchai Wongwises et.al, Simultaneous heat and mass
transfer characteristics for wavy fin-and-tube heat exchanger under dehumidifying
conditions, International Journal of Heat and Mass Transfer, 49, pp 132-143, 2006

[71] Liaquat Ali Khan, Dr. Ali El-Ghalban, Heat Exchanger Exergetic Lifecycle Cost
Optimization using Evolutionary Algorithms, WSEAS Transaction on Heat and
Mass Transfer , Issue 1, Volume 3,, pp125-136. January 2008

[72] Liaquat A. Khan, Fathi Mahfouz, Thermoeconomic Lifecycle Energy Recovery

System Optimization For Central Air-Conditioning System Using Evolutionary
Technique, Mehran University Research Journal of Engineering & Technology,
Volume 31, No. 1, pp 13-28, January, 2012.

[73] Liaquat A. Khan, Fathi Mahfouz, Thermoeconomic Lifecycle Cost Optimization

of an Annular Fin Heat Exchanger, Pakistan Journal of Engineering and Applied
Sciences, vol 11, pp 129-140, July 2012.

193