Documente Academic
Documente Profesional
Documente Cultură
Author
Thesis Supervisor
_____________________________ ______________________________
External Examiner‟s Signatures External Examiner‟s Signatures
Prof. Dr. Bsharat Ullah Malik Prof. Dr. Abdul Ghafoor
iii
Supervisor
Asst. Prof. Dr. Muhammad Shehryar
iv
Foreign Expert Evaluation Report
Gary M. Sandquist
v
Foreign Expert Evaluation Report
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)
Mikhael Gorokhovski
vi
Dedicated
To my family, my sisters
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
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
The commitment of the remaining committee members, Dr. Arshad Qureshi, Dr.
Zafar Ullah Koreshi and Dr. Shahab Khushnood also recognized. Their comments and
the continued generous support. Special thanks to Dr. Muhammad Shahid Khalil for his
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
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.
xi
Table of Contents
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 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
xvi
List of Tables
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
xvii
Nomenclature
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
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
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.
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 1.2 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]
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.
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
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
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
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
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
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.
Recombination
Population
Mutation
Termination
Offspring
Survivor selection
Figure 2 1 The general scheme of an Evolutionary algorithm flow chart
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
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.
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
2.2.2 Recombination
19
Figure 2 3 Recombination techniques dominant recombination (at the bottom) and the intermediate
recombination (at the right side) [1]
Discrete recombination
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.
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.
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.
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
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)
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.
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)
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.
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
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
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.
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)
UTLMTD F
(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)
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
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
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].
4 . (Pt 2 πd o2 /4)
d eq ; (3.12)
πd o
For triangular tube arrangement is calculated as:
To calculate tube and shell side Reynolds numbers the flow velocities are
evaluated using the following relations
mt np
vt . (3.14)
πd o
2
Nt
ρt
4
33
ms
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
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).
Qm
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
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.
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
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.
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.
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
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.
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)
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..
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
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)
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.
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)
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
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.
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].
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
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].
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
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]
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
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).
Where A is the heat transfer area of the exchanger for given heat duty.
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
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
10 ≤ m C ≤ 15
1.25 ≤ PR ≤ 1.5
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].
Do
NSTP m C kg/s L (m)
(mm)
PTT PR PLA BSR
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.
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
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 ($)
28000
26000
24000
22000
0 0.2 0.4 0.6 0.8 1 1.2
BSR
Figure 3. 7 . Comparison of AC-BSR for different pitch ratio
20000
Cost ($/Year)
10000
5000
0
Ref [13 ] This work Ref [13 ] This work Ref [13 ] This work Ref [13 ] This work Ref [13 ] This work
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.
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.
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.
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)
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])
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].
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
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)
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)
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
Ar αr
Af αf (4.20)
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
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.
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.
h'
Af A
h's (4.32)
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.
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).
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
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)
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 Cm
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)
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
(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
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.
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]
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.
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.
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.
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)
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
S gen
NS (4.60)
C max
Where S gen is the rate of entropy generation.
And
T P
ΔS m
C p ln out Rln out (4.62)
T P
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 1TL
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)
(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
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.
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.
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
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.
Discounted operating cost were computed with N = 10 years, annual discount rate
i = 10%, energy cost Cfuel = 0.12 $/kWh, daily working hours Hd =18 and number of days
in a year the system is used are Dy =180 days.
245000
240000
Net Saved Energy ($)
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
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
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.
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)
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
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.
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
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
qm
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.
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
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]
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
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)
f C f Re mf (5.11)
lnf 2 lnf 1
mf
lnRe 2 lnRe 1 (5.12)
C f e bf
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).
(d f d o )
Lf (5.14)
2
100
Ai Di
(5.15)
A 0 0.5Pf (d f d 0 ) d 0 Pf t f d 0
2 2
G Dh
Re o (5.16)
μ
Where Dh is called hydraulic diameter and it is defined as
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
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
h o Per
mf (5.28)
A o K fin
Where “Per” is the perimeter of fin surface
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)
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].
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)
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.
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].
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
ψ d (m
i Δψ i m
o Δψ o ) (5.44)
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.
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
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
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.
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
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
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:
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
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
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
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
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
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
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
152
Appendix-E
Plate-Fin Heat Exchanger Convergence Figures
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
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
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
-4
x 10
5
4.5
4
Individuals t
3.5
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
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
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
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
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;
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)
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"
P= (TCout-TCin)/(THin-TCin);
R= (THin-THout)/(TCout-TCin);
F factor function
ac= (q*1000)/(uc*dtm) ; %The surface area of the heat exchanger for the
clean condition"
af1= (q*1000)/(uf1*dtm)
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"
Tw=(1/2)*((TCin+TCout)/2+(THin+THout)/2);
%
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
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
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"
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
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,
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
167
%kC=7*(10^-5)*(TCavg)+0.0034
prH=(dvisH*CpH*1000)/kH
prC=(dvisC*CpC*1000)/kC
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 "
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"
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)
hHot = (jH*GH*CpH*1000)/((prH)^(2/3));
hCold= (jC*GC*CpC*1000)/((prC)^(2/3));
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"
Tw=(THavg+Rratio*TCavg)/(1+Rratio);
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)
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
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"
hHotr = (jHr*GHr*CpHr*1000)/((prHr)^(2/3));
hColdr = (jCr*GCr*CpCr*1000)/((prCr)^(2/3));
UAr=((1/(etaoHr*hHotr*AHr))+(1/(etaoCr*hColdr*ACr))+Rw)^(-1) ;
%"Overall heat tranfer coefficient on the hot fluid"
effecrc=1-exp((1/Cstar)*((ntur)^0.22)*(exp(-Cstar*((ntur)^0.78))-
1)) %"Effectiveness calculation based on rating problem"
172
lamdaC=(kf*AkC)/(lct*CpC*1000*mC); %"Logitudinal conduction
parameters on the cold fluid"
effectf=(THin-THoutrf)/(THin-TCin)
hcout=0.0646*(TCoutrf^2)-36.098*TCoutrf+5073.6
hcin=0.0646*(TCin^2)-36.098*TCin+5073.6
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 %% "
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;
% COST OF CROSS FLOW PALTE FIN HEAT EXCHANGER FOR ENERGY RECOVERY
SYSTEM *
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 $"
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"
[c1,c2,Qlmtd1,Qlmtd2,Bvalue] = Cvalue(Qlmtd,PH)
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));
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,
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
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
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 "
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)
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"
multf= 1/(interest+1);
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
% 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;
% 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];
Pop.individuals = allIndividuals(:,1:esopt.mue);
Pop.fit = allFitness(:,1:esopt.mue);
Pop.sigma = allSigmas(:,1:esopt.mue);
Reproduction Function
IX = ceil(esopt.mue*rand(1,esopt.PreLambda));
Pop.PreOffspring = Pop.individuals(:,IX);
Pop.PreOffspringSigma = Pop.sigma(:,IX);
187
REFERENCES
[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.
[7] Sommerer F,. Evaluation of Absorption Cycles with Respect to COP and
Economics. International Journal of Refrigeration. 19(1): pp.19-24, 1996.
[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.
[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
[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
[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.
[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.
[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
[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.
190
[40] Beyer H-G, The Theory of Evolution Strategies, National Computing Series,
Springer, Heildelberg, 2001.
[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.
[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
[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
[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.
[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
193