AMERICAN UNIVERSITY OF BEIRUT

EXPERIMENTAL AND THEORETICAL STUDY OF AN

INTEGRATED THERMOELECTRIC-PHOTOVOLTAIC
SYSTEM FOR AIR DEHUMIDIFICATION AND FRESH
WATER PRODUCTION

by
MUHYIDDINE ABED EL HAKEEM JRADI

A thesis
submitted in partial fulfillment of the requirements
for the degree of Master of Mechanical Engineering
to the Department of Mechanical Engineering
of the Faculty of Engineering and Architecture
at the American University of Beirut

Beirut, Lebanon
February 2011
 AMERICAN UNIVERSITY OF BEIRUT

EXPERIMENTAL AND THEORETICAL STUDY OF AN

INTEGRATED THERMOELECTRIC-PHOTOVOLTAIC
SYSTEM FOR AIR DEHUMIDIFICATION AND FRESH
WATER PRODUCTION

by
MUHYIDDINE ABED EL HAKEEM JRADI

Approved by:

______________________________________________________________________
Dr. Nesreen Ghaddar, Professor Advisor
Department of Mechanical Engineering

______________________________________________________________________
Dr. Kamel Ghali, Associate Professor Member of Committee
Mechanical Engineering

______________________________________________________________________
Dr. Fouad Azizi, Assistant Professor Member of Committee
Mechanical Engineering

Date of thesis defense: February 10, 2011

 AMERICAN UNIVERSITY OF BEIRUT

THESIS RELEASE FORM

I, Muhyiddine Abed el Hakeem Jradi

authorize the American University of Beirut to supply copies of my thesis project

to libraries or individuals upon request.

do not authorize the American University of Beirut to supply copies of my thesis

project to libraries or individuals for a period of two years starting with the date of
the thesis project defense.

____________________
Signature

____________________
Date
 ACKNOWLEDGMENTS

First and foremost, I would like to express my deep and sincere gratitude to my
advisors at the American University of Beirut for their continuous help and support that
enabled me to accomplish this work.

The guidance, understanding and encouraging of my advisors Prof. Nesreen

Ghaddar and Dr. Kamel Ghali was very essential to initiate and boost this work. Their
motivation and follow up was very helpful for me to solve all problems I faced in my
research during the period spent on accomplishing my master’s degree at AUB. This
thesis would not have been completed without their advice and unfailing patience. It
was a most rewarding learning experience to work under their guidance.

I would like to acknowledge with great appreciation Dr. Fouad Azizi for his
constant advice, constructive criticism, valuable notes and comments and for reviewing
my thesis as a member of my thesis committee.

I would like to acknowledge the support and help of the faculty at the
Mechanical Engineering Department for enhancing my scientific knowledge, analytical
and learning skills.

I would like to thank the staff at the American University of Beirut for their
continuous help and technical support.

During this work I have worked and collaborated with many friends and
colleagues in the Applied Energy program for whom I have great regard.

Most importantly, none of this would have been possible without the love and
patience of my family. My family, to whom this thesis is dedicated, has been a constant
source of love, concern, support and strength all these years. I would like to express my
heart-felt gratitude to my mother, father, and lovely brothers and sisters. I love you all.

v
 AN ABSTRACT OF THE THESIS OF

Muhyiddine Abed El Hakeem Jradi for Master of Mechanical Engineering

Major: Applied Energy

Title: Experimental and Theoretical Study of an Integrated Thermoelectric-Photovoltaic

System for Air Dehumidification and Fresh Water Production

The main objective of this work is to present an integrated thermoelectric-

photovoltaic renewable system to dehumidify air and produce fresh water. The system
is combined with a solar distiller humidifying ambient air to enhance distillate output
and system performance to meet the specified fresh water needs for a residential
application. The presented system is a totally renewable energy-based system taking
advantage of the clean solar energy to feed the thermoelectric modules with electricity
through the photovoltaic panels and to increase the water content of air and enhance the
water production through the solar distiller.

A model is developed to simulate the air dehumidification process using

thermoelectrically cooled channels. The model predicts the temperature and the
humidity of the air leaving a cooled channel, the variation of the air temperature through
the channel, the TEC modules hot and cold side temperatures and their corresponding
heating and cooling capacities as well as the electric power consumption and the
amount of collected water condensate. Experiments were performed to validate the
developed model results. It was found that the model of the thermoelectrically cooled
channel predicts well the variation in the air temperature through the channel with a
maximum relative error in air temperature less than 2.4 %. In addition, the simulation
model predicts well the amount of water condensate produced by the integrated system
with a maximum relative error of 8.3 %.

An optimization problem is formulated to design and set the integrated system

optimal operation to produce 10 liters of fresh water per day meeting the fresh water
needs of a typical residential. Using five TEC channels of length of 1.2 m and area of
0.07×0.05 m2 integrated with 1.2 m2 solar distiller that recirculates a constant air mass
flow rate of 0.15 kg/s is capable of meeting water demand when air mass flow rate
through each TEC channel is optimally set at 0.0155 kg/s. The associated optimal
electrical current input to the TEC modules varied depending on the month and is set at
2.2 A in June, 2.1 A in July and 2.0 A in August, September and October. It was found
also that the optimal operation total energy consumption varied between 5.32 kWh/day
in October and 6.73 kWh/day in June, with about 17.26 % and 45.25% saving
respectively compared to the conventional vapor compression systems.

vi
 CONTENTS

ACKNOWLEDGMENTS………………………………………. v

ABSTRACT………………………………………..…………... vi

LIST OF ILLUSTRATIONS………………………..…………...x

LIST OF TABLES………………………………………....…...xiii

LIST OF NOMENCLATURE…………………………….........xiv

Chapter

1. INTRODUCTION……………………………………………..1

2. THERMOELECTRIC COOLING TECHNOLOGY…….…....5

2.1. Thermoelectricity Module……………………………………………………...5

2.2. Thermoelectric Basics………………………………………………………….9

2.3. Thermoelectric Cooling……………………………………..………………...10

3. LITERATURE REVIEW…………………………………….12

4. PROBLEM STATEMENT AND SYSTEM DESCRIPTION. 16

4.1. Thermoelectric Cooling Channel……………………………………………...17

4.2. Solar Photovoltaic Modules…………………………………………………...17

4.3. Solar Distiller Humidifying Unit……………………………………………...18

4.4. System Components Integration……………………………………………....18

5. MATHEMATICAL FORMULATION……………………... 21

5.1. Thermoelectric Cooling Module……………………………………………... 21

vii
 5.2. Cooling Channel……………………………………………………………... 26

6. NUMERICAL SOLUTION METHODOLOGY……………. 29

7. EXPERIMENTAL METHODOLOGY……………………... 33

7.1. Experiment Description……………………………………………………… 33

7.1.1. Climatic Chamber…………………………………………………… 35

7.1.2. Thermoelectric Cooling Modules…………………………………… 36
7.1.3. Heat Sinks…………………………………………………………… 39
7.1.4. Electric Power Supplies……………………………………………... 40
7.1.5. Heater and Humidifier………………………………………………. 40
7.1.6. Axial Air Fan………………………………………………………... 40
7.1.7. Measurements and Recordings……………………………………… 41

7.2. Experimental Protocol………………………………………………………...42

7.3. Experimental Setup…………………………………………………………... 43

7.4. Experimental Results………………………………………………………… 45

7.4.1. Experiment Session 1……………………………………………….. 45

7.4.2. Experiment Session 2……………………………………………….. 47
7.4.3. Experiment Session 3……………………………………………….. 50
7.4.4. Experiment Session 4……………………………………………….. 52

7.5. Simulation Model Validation………………………………………………… 55

8. CASE STUDY………………………………………………. 59

9. RESULTS AND DISCUSSION…………………………….. 64

10. CONCLUSION AND RECOMMENDATIONS ………......73

Appendix

1. THE SOLAR STILL MODEL………………………………. 76

2. HOURLY FRESH WATER PRODUCTION DATA………..80

viii
3. HOURLY COOLING CAPACITY DATA…………………. 83

4. HOURLY HEAT DISSIPATION DATA…………………… 86

5. HOURLY ENERGY CONSUMPTION DATA…………….. 89

6. TE-PV SYSTEM DATASHEETS………..…………………. 92

REFERENCES…………………………………………………. 97

ix
 ILLUSTRATIONS

Figure Page

1. Thermoelectric cooling module……………………………………………….... 6

2. P-type semiconductor pellet operation………………………………………….. 7

3. N-type semiconductor pellet operation…………………………………………..7

4. General sketch for the integrated system units………………………………… 16

5. Thermoelectric cooling module operation……………………………………...22

6. Thermoelectric cooling module energy modeling……………………………...23

7. Iterative method flow chart……………………………………………………..31

8. Experimental system unit……………………………………………………… 34

9. HTS700B climatic chamber with the HTS700 temperature and relative

humidity control system……………………………………………………….. 36
10. TEC1-12714 geometric characteristics………………………………………... 37

11. The 55*55 mm HXHF CPU coolers used with the heat sinks and the DC fans. 39

12. The built thermoelectric cooling channel……………………………………… 43

13. The electric power supplies used in the experimentations…………………….. 44

14. Amount of water condensate versus the input electric current…………………49

15. Air temperature variation with electric current input………………………….. 49

16. Amount of water condensate versus the air mass flow rate…………………… 51

17. Air temperature variation with air mass flow rate…………………………….. 52

18. Amount of water condensate versus the input air temperature………………... 54

19. Air temperature variation with air mass flow rate……………………………... 55

x
20. Plots of (a) water condensate versus electric current as predicted by the model
and recorded experimentally, (b) water condensate versus mass flow rate as
predicted by the model and recorded experimentally and (c) water condensate
versus inlet air temperature as predicted by the model and recorded
experimentally…………………………………………………………………. 57
21. Air temperature variation through the cooling channel for different electric
current inputs, 0.01 kg/s air mass flow rate, 28 °C ambient temperature and 30
°C inlet air temperature (model predictions-experimental data comparison)…..58
22. Variation of water condensate with the variation of air mass flow rate under
different electric current supply to the system in August……………………… 61
23. The amount of fresh water produced and the PV power consumption as a
function of electric current input to the TEC modules for (a) August and (b)
September……………………………………………………………………… 63
24. Hourly fresh water production………………………………………………….66

25. Daily fresh water production for the five months………………………………67

26. Total fresh water production over 10 hours of the PV-powered TEC system
with and without the solar distiller…………………………………………….. 68
27. Plots of (a) the cooling capacity pumped by the thermoelectric cooling
modules/day and (b) the associated system energy consumption from the PV
in kWh/day over the summer months…………………………………………..69
28. The cost of fresh water production by the system for five months compared
to the cost for the conventional vapor compression systems…………………...71
29. Solar still sketch………………………………………………………………...77

30. Hourly fresh water production in June………………………………………… 80

31. Hourly fresh water production in July... ………………………………………81

32. Hourly fresh water production in August……………………………………… 81

33. Hourly fresh water production in September………………………………….. 82

34. Hourly fresh water production in October……………………………………...82

35. Hourly cooling capacity pumped by the TEC modules in June……………….. 83

36. Hourly cooling capacity pumped by the TEC modules in July………………... 84

37. Hourly cooling capacity pumped by the TEC modules in August…………….. 84

38. Hourly cooling capacity pumped by the TEC modules in September………… 85

xi
39. Hourly cooling capacity pumped by the TEC modules in October…………….85

40. Hourly heat dissipated by the TE cooling system in June……………………... 86

41. Hourly heat dissipated by the TE cooling system in July………………………87

42. Hourly heat dissipated by the TE cooling system in August…………………...87

43. Hourly heat dissipated by the TE cooling system in September………………. 88

44. Hourly heat dissipated by the TE cooling system in October…………………. 88

45. Hourly system energy consumption in June…………………………………… 89

46. Hourly system energy consumption in July…………………………………… 90

47. Hourly system energy consumption in August…………………………………90

48. Hourly system energy consumption in September…………………………….. 91

49. Hourly system energy consumption in October……………………………….. 91

50. Qc vs Voltage under various ΔT (Th = 27°C)………………………………..... 92

51. Qc vs ΔT under various currents (Th = 27°C)………………………………..... 93

52. Voltage vs ΔT under various currents (Th = 27°C)…………………………..... 93

53. Qc vs Voltage under various ΔT (Th = 50°C)………………………………..... 94

54. Qc vs ΔT under various currents (Th = 50°C)………………………………..... 94

55. Voltage vs ΔT under various currents (Th = 50°C)…………………………..... 95

56. UE 200 M5 electrical specifications…………………………............................96

57. UE 200 M5 mechanical specifications…………………………........................ 96

xii
 TABLES

Table Page

1. Performance specifications for the (n) and (p) semiconductor elements……… 38

2. TEC1-12714 performance specifications……………………………………… 38

3. Experimental results of the overall hot side thermal resistance……………….. 46

4. Water condensate and air temperature variation for different input electric
current…………………………………………………………………………..48
5. Water condensate and air temperature variation for differentt input air mass
flow rate ..............................................................................................................50
6. Water condensate and air temperature variation for different ambient and
input air temperature ……………………………………………………...........53
7. Hourly and total amount of condensed water over the five months ……….......65

xiii
 NOMENCLATURE

A Area, m2

CCD charge coupled device

COP coefficient of performance

Cp specific heat coefficient at constant pressure, J/kgK

Cpg glass specific heat coefficient, J/kg.K

Cpv water vapor specific heat coefficient, J/kg.K

Cp_w water vapor specific heat capacity, J/kgK

DC direct current

DNA deoxyribonucleic acid

Gf geometry factor, cm

h enthalpy, J/kg

hfg latent heat of vaporization, J/kg

hi internal heat convective transfer coefficient, W/m2K

I electric current, A

I(t) solar radiation, W/m2

k thermal conductance of the module, W/K

LED light emitting diode

m mass, kg

mass flow rate, kg/s

Nb number of thermocouples

P electric power, W

xiv
PV photovoltaic

Q thermoelectric net heat, W

R thermal resistance, K/W

Re electrical resistivity, ohm

RH relative humidity, (%)

t time, s

T temperature, °C

U voltage difference, V

V velocity, m/s

W humidity ratio, kg H2O/kg of air

X salinity concentration

ΔT temperature difference, °C

Greek letters

 seebeck coefficient, V/K

 transmissivity

 emissivity

 stephan boltzman constant

Subscripts

A air

amb ambient

C cold

c-a convection with the ambient air

xv
c-g air convection with glass

c-w air convection with water

cond condensate

dist distiller

dry dry conditions

e thermoelectric element

g glass

h hot

i internal

in input

m module

max maximum

min minimum

n n type semiconductor

out output

p p type semiconductor

r1 radiation (water and glass)

r2 radiation with ambient

sky sky

w water

wet wet conditions

xvi
 Dedication addressed

To my beloved Mother, Father,

brothers Mohammed and Majed

and sisters Farah and Malak

xvii
 CHAPTER 1

INTRODUCTION

In spite of the technological and the industrial advancement accompanied with

the fast rates of urbanization, the lack of fresh water resources emerges as one of the

critical problems threatening the humanity and the nature around us. The decline in the

fresh water availability and quality is affecting millions of people especially in the

developing and poor countries. Recent Statistics show that about 15 % of the people in

the Arab World don’t have access to clean and fresh water, and this percentage is

expected to increase in the upcoming years [1]. The warm and dry climates in the

Middle East and North Africa exhibit the severest fresh water scarcity problem due to

the increase in the salinity rate of the ground water and the sporadic rainfall [2].

According to Miller et al. [1] water scarcity and fresh water shortages affect 88

developing and poor countries contributing to about half the world’s population. In

2025, the number of people suffering from fresh water scarcity and the lack of access to

clean water resources is expected to be multiplied by four. In these countries, people’s

need for fresh water resources is faced with their limited financial conditions and their

miserable life situations, so the development of low cost efficient desalination

techniques will help in solving part of this rigorous problem.

For more than fifty years, many desalination techniques were developed to

solve part of this impending water shortage problem and meet the fresh water

requirements. The main focus was to enhance the extraction of fresh water from sea

water or other salty water sources. Common desalination techniques in this field include

Vapor Compression, Membrane Distillation, Electrodialysis, Reverse Osmosis, Multi

1
Effect Evaporator, Multi Stage Flash Distillation and others. However, these

desalination techniques are energy-intensive processes and are characterized by high

conventional energy consumption rates. Due to their high consumption of fossil fuels,

these desalination processes are usually accompanied with harmful environmental

effects and nature unbalances. Moreover, they are in need for regular and complex

maintenance with high operation costs. On the other hand, the huge unprecedented

increase in oil prices and conventional energy resources makes the traditional

desalination processes economically very expensive with the need for huge energy

capital costs.

All of this raises the need for advanced, energy efficient and environmental

friendly solutions including fresh water extraction from the atmosphere while utilizing

renewable energy resources. Using photovoltaic-powered thermoelectrically cooled

surfaces to extract water from the atmosphere is one viable solution that takes advantage

of the humid climate prevailing in many areas of the Middle East. In the days with

relatively low air humidity, the power consumption of the thermoelectric cooling

modules would drastically increase which means that other methods are to be sought to

humidify ambient air. Integrating a solar distiller with a thermoelectric cooling system

would increase the water content of the air into the cooling system to produce the

desired fresh water at lower electrical power consumption.

The aim of the thesis is to design a solar-driven thermoelectric cooling system

integrated with a solar distiller humidifying unit for air dehumidification and fresh water

production. The work concentrates on studying the viability of applying the advanced

thermoelectric technology in designing a low-power dehumidifying device, which can

be driven using renewable energy source, the solar PV panels, to generate water in the

2
Lebanese humid climate. In this work, we will present an experimental and theoretical

study of a new low-power dehumidifier made up of thermoelectric cooling modules for

air dehumidification and fresh water generation. Thermoelectric cooling modules are

emerging today as one of the most promising alternatives to conventional energy

resources and systems in cooling and refrigeration applications. This is due to their

compact, maintenance-free, gravity dependent, durable, acoustically silent and reliable

design and operation.

The proposed thermoelectric dehumidifier is electrically driven by solar energy

through the photovoltaic solar panels. Therefore, with the free electrical energy supply

and the absence of any chlorofluorocarbons or harmful gases, our proposed

thermoelectric-photovoltaic dehumidification system will present itself as an efficient

and environmental friendly system consuming less energy, conserving the environment

and producing the desired fresh water. In addition, the thermoelectric-photovoltaic

proposed system will be integrated with the solar distiller humidifying unit to enhance

the fresh water production by increasing the water content of the air exiting the solar

still. This humid air will be introduced to the cooling channel to be dehumidified using

the thermoelectric cooling modules which in turn are supplied with the needed electrical

current by the photovoltaic solar panels. All of this makes the overall system as a whole

a totally renewable energy-based system without any vapor compression cycle or any

conventional energy resource. The thesis main objectives can be summarized as

follows:

 Present a totally renewable energy-system for air dehumidification and fresh

water production based on the thermoelectric and the photovoltaic techniques.

3
  Develop a robust simulation model of the thermoelectric cooling channel to

simulate the dehumidification of air through the thermoelectric channel and predict the

thermal and the electrical performance of the modules as well as the amount of

condensate generated.

 Design and build an experimental setup for the proposed system to validate the

numerical model results and predictions regarding the water condensate collected and

the various system performance parameters.

 Apply the integrated proposed system to a case study in Lebanon to study the

feasibility of using such systems to generate fresh water in the Lebanese humid weather.

4
 CHAPTER 2

THERMOELECTRIC COOLING TECHNOLOGY

The thermoelectric module can act either as a heat pump or as a power

generator. Applying a DC voltage difference across the thermoelectric module, an

electric current will pass through the module and heat will be absorbed from one side

and released at the opposite side. On the other hand, maintaining a temperature

difference between the two junctions of the module, a voltage difference will be

generated across the module and an electrical power is delivered.

Thermoelectric modules’ applications cover a wide spectrum of product areas.

This includes commercial products such as beverage coolers, portable picnic cooling

boxes, drain fountain cooling and compact refrigerators. In addition, thermoelectric

modules are used in laser diodes, integrated circuits and microprocessor cooling, CCD

and LED cameras, infrared detectors cooling as well as night vision equipment for

military applications. Thermoelectric modules are also very effective in dew point

hygrometer, oil pour point apparatus, vidicon tubes, blood analyzers and recently well

known in DNA research. Moreover, thermoelectric modules are getting step by step on

the road of air conditioning, on-board refrigeration systems, compact heat exchangers,

immersion coolers as well as dehumidifiers and there are a lot of innovations in those

fields.

2.1. Thermoelectric Module

A thermoelectric module typically consists of a number of integrated

thermoelectric elements connected electrically in series and thermally in parallel. There

5
are a lot of materials used in producing the thermoelectric elements; however, the most

abundant and popular one is Bismuth Telluride.

Fig. 1: Thermoelectric cooling module

We have two types of semiconductor pellets: P-type and N-types. A P-type

semiconductor pellet is positively doped so that the charge carriers in the material are

positive. Those positive charge carriers, called holes, enhance the flow of electrons

through the material under an applied voltage. They are attracted by the negative pole of

the DC power supply, so the holes flow in the opposite direction to the flow of the

electrons in the circuit.

6
 Fig. 2: P-type semiconductor pellet operation

On the other hand, N-type semiconductor pellet is negatively charged so that

the charge carriers that are negative (called electrons), are attracted by the positive pole

of the DC power supply.

Fig. 3: N-type semiconductor pellet operation

7
 As shown in the above two figures, the flow of heat follows the flow of the

holes in the P-type semiconductor, and it follows the flow of the electrons in the N-type

semiconductors. Therefore, the electrical current moves alternately back and forth

between the two substrates through each N-type and P-type thermoelectric

semiconductor element. As a result, having a couple of semiconductors formed of N-

type and P-type semiconductor with a junction between them is the most appropriate

way in configuring a series circuit allowing to keep the heat flow in the same direction.

In order to provide a platform for the semiconductors and the small tabs

connecting them, metalized ceramic substrates are used in general to hold everything

together. Generally, those ceramic substrates are formed of aluminum oxide (Al2O3).

They provide the mechanical and structural integrity of the module, electrically insulate

the thermoelectric elements from any external conductor surface and provide a flat and

parallel surface on both sides facilitating the mounting procedure.

The thermoelectric module performance depends mainly on three physical

characteristics, the Seebeck coefficient, the thermal conductivity and the electric

resistivity of the thermoelectric module. Those characteristics are temperature

dependent and vary according to the operating temperature. Moreover, the temperature

difference between the cold and the hot sides of the thermoelectric module is a crucial

factor in the operation of the module; the higher the temperature difference across the

hot and the cold junctions, the lower the coefficient of performance of the

thermoelectric module.

8
2.2. Thermoelectricity Basics

In the beginning of the nineteenth century, two scientists, Thomas Seebeck and

Jean Peltier, discovered the thermoelectric phenomenon which constitutes the basis of

all the thermoelectric applications and industry nowadays. The phenomenon is based on

three main effects: the Seebeck effect, the Thompson effect and the Peltier effect.

Thermoelectric power supply generators are based on the Seebeck phenomenon

which is based on voltage generation along a conductor subjected to a gradient of

temperature. Thus when a temperature gradient is applied to a conductor, an

electromotive force is produced. The voltage difference generated is proportional to the

temperature difference across the thermoelectric module between the two junctions, the

hot and the cold one. The Seebeck coefficient is defined as the ratio of the temperature

gradient to the voltage difference.

The Thompson effect governs the cooling and the heating of a material

carrying a current and subjected to a temperature gradient. It states that any conductive

material across which a temperature gradient is applied will either emit or absorb heat

based on the type and the properties of the material.

The Peltier effect is the main contributor to all thermoelectric cooling

applications. It is the responsible for heat removal and heat absorbance. It states that

when an electric current flows across two dissimilar conductors, the junction of the

conductors will either absorb or emit heat depending on the flow of the electric current.

The heat absorbed or released at the junction is proportional to the input electric current.

The constant of proportionality is called the Peltier coefficient.

9
2.3. Thermoelectric Cooling

A thermoelectric cooling module, or known as a thermoelectric cooler, is a

simple solid-state semiconductor-based electronic device that functions as a heat pump.

A thermoelectric cooler simply can convert the input electrical energy into thermal

energy building up a temperature gradient between its hot and cold sides. Thus, by

applying a DC voltage across the thermoelectric cooling module, heat will be absorbed

from the cold side and released at the opposite hot side [3].

Forty years ago, thermoelectric coolers were introduced as an ideal solution in

many small-scale cooling applications including military, aerospace, medical, optical,

industrial and commercial products cooling. In addition, small thermoelectric

refrigerators were designed and have proven to be applicable in many fields. Moreover,

thermoelectric cooling modules could be supplied with electrical current produced by

the photovoltaic solar arrays. Thus, using solar energy to power the cooling modules is

an energy efficient and environmental friendly solution to meet the cooling and the

refrigeration demands.

Thermoelectric cooling is not the best cooling application in all fields;

however, thermoelectric modules provide a solution in many cooling applications

especially in many problems regarding low amount of heat removal. Some of the

advantages of thermometric cooling include:

 Small size and very light weight compared to traditional mechanical systems

 Compact and reliable due to their solid-state construction

 Can reach steady-state operation very quickly

 Have no moving parts, fluids, any compressors or bulky pipes

 Durable and maintenance-free with very long operation life

10
 Ability to heat and cool with the same module by polarity inversion

 Ability to cool to very low temperature below zero degrees Celsius

 Can work in any orientation including zero gravity environments

 Ability for temperature control with a closed-loop temperature control circuit

 Acoustically silent and generate no electrical noise

 Very effective in spot cooling to cool specific areas or components

 Environmentally friendly without chlorofluorocarbons in their operation

 Operate by DC power and can be powered directly by PV solar cells.

11
 CHAPTER 3

LITERATURE REVIEW

Considerable amount of research has been conducted in the field of

thermoelectric cooling to predict and optimize the performance of thermoelectric

modules [4 -10]. Huang et al. [4] developed a new design method for the thermoelectric

cooling modules. This was based on experimentally testing the modules in order to

develop performance curves and empirical relations which were used later in the design

and calculations. They studied the effect of the heat sink on the overall performance of

the cooling module and reported that the two major factors for enhancing the efficiency

of the thermoelectric modules are the heat sink thermal resistance and the appropriate

cooling capacity to reach the optimal performance. Similar work reported by Chang et

al. [11] who studied the effect of the current input and the heat load on the performance

of the thermoelectric modules. They found that for a given heat load, there is a

corresponding optimal current for the module, and that the thermoelectric modules

become more effective at lower heat loads.

Cosnier et al. [12] presented an experimental and analytical study of a

thermoelectric system for air pre-cooling and pre-heating and found that the system is

feasible and efficient reaching a coefficient of performance (COP) of 1.5 in the cooling

mode and about 2 in the heating mode. This is achieved by supplying an optimal electric

current to the modules and maintaining a desirable temperature difference of 10 °C

using a proper heat sink. However, these results were not attained in the experimental

work due to the high resistance of the heat sink used. The work of Lertsatitthanakoran et

al. [13] has concentrated on studying the thermal comfort of a thermoelectric ceiling

12
cooling panel system. They used 36 cooling modules with aluminum at the cold side

and water heat exchanger at the hot side. A cooling capacity of 201.6 W was attained

along with a COP of 0.82 under 1 A electric current input to the thermoelectric

modules.

Furthermore, several studies investigated the viability of using photovoltaic

(PV) solar panels to supply electricity to the thermoelectric cooling modules [14-15].

Gillot et al. [14] used thermoelectric modules for small-scale space conditioning and

built a system prototype that consisted of eight cooling modules to generate a cooling

capacity up to 220 W with a COP of 0.46 under current input of 4.6A. They reported

that the hot side temperature is proportional to the input current and that the applications

with lower current input will have higher COP but lower cooling capacity. In addition,

they found that the use of solar PV to feed the thermoelectric modules would be

competitive and feasible if the PV's price is lower than £1.25/W. La Pierre et al. [15]

studied the use of thermoelectric modules supplied with electrical power generated by

photovoltaic panels for air pre-cooling and pre-heating. They investigated the optimal

number of thermoelectric modules that could be used as well as the optimal electric

current to the thermoelectric modules under different radiation intensities and found that

beyond those optimal conditions the system COP will begin to decline. Their study

shows that thermoelectric cooling could be efficient for summer cooling for small-scale

applications; however, the high investment cost of the PV panels should always be

taken into account.

Moreover, many researchers concentrated on investigating the performance of

the thermoelectric cooling modules in refrigeration applications. Dai et al. [16]

investigated the viability of building a thermoelectric refrigerator driven by solar PV

13
cells to serve people in remote and rural areas. Their proposed system was tested under

different radiation intensities, and they found that the system performance is mainly

dominated by the available intensity of irradiation. The refrigerator temperature ranges

between 5 and 10C and the COP is about 0.3. Another portable solar thermoelectric

refrigerator was designed by Abdul-Wahab et al. [17]. The results presented in their

work shows that the refrigerator is able to attain a minimum temperature of 5C with a

COP of 0.16. They stated that the efficiency of the thermoelectric refrigerators is still

lower than its compressor counterparts, but some improvements were proposed to

increase the COP such as improving the thermal interfaces between the modules and the

heat sinks as well as decreasing the thermal resistance of the sinks.

As shown from the literature above, the majority of the studies and work done

concentrated on cooling and refrigeration using thermoelectric modules. However,

thermoelectric dehumidification was examined by very few researchers. One study was

conducted by Vian et al. [18] who developed a small dehumidifier prototype made up of

three thermoelectric cooling modules and a computational model (AERO) was used for

the design and the optimization of the modules’ performance. Their system consumed

100 W at optimum conditions of 27C and 82% relative humidity with a COP much

lower than the vapor-compression devices.

The aim of this study is to design a solar-driven thermoelectric cooling system

integrated with solar distiller for air dehumidification and fresh water production. The

work will concentrate on studying the viability of applying thermoelectric technology in

designing a low-power dehumidifying device driven by solar PV panels to generate

fresh water in the Lebanese coastal humid climate. A robust model will be developed to

simulate the dehumidification of the humid air through the thermoelectrically cooled

14
channel surface and to predict the thermal and the electrical performance of the modules

as well as the amount of condensate generated. A TEC channel prototype will be built

and experiments will be carried to validate the model predictions of the amount of water

condensate collected and the other system performance parameters. A case study will be

presented to assess the integrated dehumidification system in achieving the objectives of

meeting the desired fresh water needs in the Lebanese humid climate.

15
 CHAPTER 4

PROBLEM STATEMENT AND SYSTEM DESCRIPTION

Fig. 4 depicts the PV-powered thermoelectric cooled channel integrated with a

solar distiller. The system is composed of three major elements, the thermoelectric

cooled channel/s, the photovoltaic solar panels, and the solar distiller humidification

unit.

Fig. 4: General sketch for the integrated system units

16
4.1. Thermoelectric Cooling Channel

It is in fact the main component in our system and it is the responsible for air

cooling, dehumidification and fresh water generation. A cooling channel will be

simulated and designed where the air is introduced to be cooled and dehumidified in this

channel. The cooling channel is sized such that one side of the channel is cooled by

thermoelectric modules situated in an adjacent vertical manner, and the other three sides

are assumed well insulated.

Those thermoelectric cooling modules will be fed directly by a DC electric

power supply, and they will play the role of heat pumps. Thus applying a voltage

difference across the thermoelectric module will result into a temperature gradient

between the two sides of the module. As a result, the module will be extracting heat

from the air flowing along the channel at the cold side of the thermoelectric cooling

modules and rejecting this heat at the other side of the modules.

4.2. Solar Photovoltaic Modules

The solar photovoltaic panels will serve as a renewable electricity supply

source. Photovoltaic generation using solar cells produce electricity directly from the

electromagnetic radiation. When the photovoltaic module is illuminated and bombarded

with photons from sunlight, a voltage difference will appear and as a result an electric

current will flow through the circuit and the connected load. Then the magnitude of

electric current produced is directly proportional to the amount of solar photons striking

the surface.

Thus, photovoltaic modules simply convert sunlight into electrical energy.

And since the equipped thermoelectric cooling modules need a DC electric power

17
supply source to operate, then the photovoltaic modules make the best appropriate

renewable solution for feeding the thermoelectric modules with the sufficient electric

power needed to pump the desired cooling capacity. It should be noted that the electrical

energy produced by the photovoltaic modules vary considerably between one month

and another depending on the solar irradiation presented and the ambient air conditions.

4.3. Solar Distiller Humidifying Unit

Since our system is proposed to produce fresh water in humid areas, then we

are in need for a humid air source to supply the thermoelectric cooling channel with the

needed air input. This role is achieved through using a solar distiller humidifying unit. A

solar distiller is a basin with a shallow depth of salty or impure water with a transparent

and vapor-tight cover sloped towards a collection channel. Therefore the solar distiller

will take advantage of the solar energy to heat and evaporate the water in the basin and

hence increase the water content of the air flowing through the distiller (as shown in

Fig. 4) and entering the cooling channel in order to enhance the fresh water production.

4.4. System Components Integration

The three components presented above will be integrated to reach the project

objective. Solar energy from the sun will be captured by the photovoltaic modules

which will generate electricity and feed the thermoelectric modules, situated along the

entire cooling channel, with the needed DC voltage. In addition, solar energy will be

used also to heat the water basin inside the solar still. Then, air from the ambient

environment will flow through the distiller from one end above the water and leaves

from the other end with higher water content. Therefore ambient air is humidified by the

18
solar distiller and then dehumidified by the thermoelectric cooling modules to extract

water condensate and generate the needed fresh water. The main parameters governing

the integrated system operation are:

 Electric current feed to the thermoelectric module (I).

 Mass flow rate of the supply air to the thermoelectric cooling channel ( ̇ ).

 Temperature and humidity ratio of the air entering the thermoelectric cooling

channel (Ta-in and wa-in).

 Ambient air conditions controlling the operation of the system (Tamb and wamb).

 Output air temperature and humidity ratio from the distiller (Ta-dist and wa-dist).

The thesis work will be divided into two main parts. In the first part, we will

develop a well robust simulation model of the thermoelectric cooling channel to

simulate the dehumidification of the humid air through the thermoelectric channel and

predict the thermal and the electrical performance of the modules as well as the amount

of condensate generated in the Lebanese climatic conditions. The model will predict the

temperature and the humidity of the air leaving the cooling channel, the variation of the

air temperature through the whole channel, the thermoelectric modules hot and cold side

temperatures and their corresponding heating and cooling capacities as well as the

electric power consumption of the modules and the amount of water condensate

produced.

This developed thermoelectric cooling channel model will be integrated with

the solar still model developed by Alsaidi et al. [19] at AUB, which predicts well the

exit air temperature and relative humidity from the distiller, in order to get benefit and

to make use of the humid air exiting the distiller as a humid source input to the

thermoelectric cooling channel. In our application, the air input to the solar distiller

19
model is not the ambient air, but it is a mixed of fresh air from outside and return air

from the distiller (as shown in Fig. 4). This will enhance the operation of the solar

distiller and will allow us to get benefit of the re-circulated air to increase the water

content of the exiting air from the distiller, therefore increasing the amount of fresh

water produced.

And in the second part of our work, we will design and build an experimental

setup for the proposed system to validate the numerical model results and predictions

regarding the water condensate collected and the various system performance

parameters. The experiments will be conducted in the Energy Lab in the SRB building

at AUB, and the results will be recorded and compared with the numerical model

predictions under the same operational and ambient conditions in order to validate the

results of the simulation model developed.

20
 CHAPTER 5

MATHEMATICAL FORMULATION

In the first part of the methodology section, the thermoelectric cooling channel

air and moisture thermal transport model is presented. This will be followed by a

description of the experimental methodology to validate the model predictions of the

water condensate collected and other system performance parameters.

5.1. Thermoelectric Cooling Module

A thermoelectric cooling module is characterized by four parameters

describing its maximal performance which are:

 ΔTmax: is the maximum temperature difference between the hot side of the

thermoelectric module and the cold side

 Qmax: is the maximal pumping cooling capacity at the cold side corresponding to

ΔT = 0

 Umax: is the maximum voltage difference that can be applied to the

thermoelectric module with the absence of any heat load

 Imax: is the maximal operational input current to the thermoelectric module at

ΔTmax

21
 Fig. 5: Thermoelectric cooling module operation

In addition, the basic operational characteristics of a thermoelectric module

include the cold side temperature (Tc), the hot side temperature (Th), the heat pumping

capacity at the cold side (Qc), the heat dissipated at the hot side (Qh), the electric current

(I) and the voltage difference (U).

22
 Fig. 6: Thermoelectric cooling module energy modeling

In our study, each thermoelectric cooling module is treated as a lumped surface

represented by uniform value for each of its operational parameters, The modeling of

the thermoelectric cooling module will follow the derivations of Tsai et al. [20], where

the operation and the performance of the module is governed by three main energy

processes which are, the Joule’s effect, the Peltier cooling effect and the thermal

conduction.

The Joule’s effect is the heat dissipated due to the additional resistive heating

of the thermoelectric elements in the module caused by the flow of electric current and

is given by

Qjoule = I2.Re (1)

23
where I is the electric current in A and Re is the electrical resistivity of the module in

ohm.

The Peltier effect describes the heat absorption or dissipation due to the

presence of two dissimilar semiconductors when an electric current flow through them.

It is given by

QPeltier = α.I.Tc/h (2)

where α is the Seebeck coefficient of the module in V/K and Tc/h is the cold or hot side

temperature.

In addition, the thermal conduction is based on the Fourier principle and is

characterized by the thermal conductivity k and given by

QCond = -k.ΔT (3)

where k is the thermal conductance of the TEC module in W/K and ΔT is the

thermoelectric module temperature difference.

The heat cooling capacity at the cold side and the heat dissipated at the hot side

can be obtained by applying steady state energy balances for both sides of the

thermoelectric cooling module [21]. As a result, the heat pumping cooling capacity at

the cold side of the thermoelectric cooling module is given by

Qc = α.I.Tc – 0.5.I2Re – k.(Th - Tc) (4)

The heat dissipated at the hot side of the thermoelectric module is given by

Qh = α.I.Th + 0.5.I2Re – k.(Th - Tc) (5)

This heat dissipated is the sum of both, the heat pumping capacity at the cold side of the

module and the electric power input to the module. In addition, the input voltage

difference applied to the thermoelectric module is given by

U = α.(Th - Tc) + I.Re (6)

24
The electric power consumption of the thermoelectric module is given by

P = U.I = Qh – Qc = α.I.(Th - Tc) + I2.Re (7)

The coefficient of performance (COP) is defined as the cooling heat capacity pumped

by the thermoelectric module divided by the electric power consumption of the module

and given by

Qc
COP = (8)
P

The three characteristics of the thermoelectric modules, the Seebeck coefficient

(α) in V/K, the thermal conductance (k) in W/K and the electrical resistivity (Re) in ohm

are given in terms of the characteristics of one thermoelectric element by the following

equations:

α = α_e Nb 2 (9.a)

Re = (Re_e Nb 2)/Gf (9.b)

k = k_e Nb 2 Gf (9.c)

where Nb is the number of thermo-couples in the thermoelectric cooling module, Gf is

the geometry factor of the thermoelectric element, α_e is the Seebeck coefficient for one

thermoelectric element in V/K, Re_e is the electrical resistivity of a thermoelectric

element in W/K and k_e is the thermal conductance for one element in W/K.

The Seebeck coefficient, the electrical resistivity and the thermal conductivity for a

single thermoelectric element can be given by the correlations presented by Gillot et al.

[14] for Bismuth Telluride thermoelectric cooling modules as:

α_e = {-0.45263 + 0.01424 Tm - 2.02517 10-5 Tm2} (10-4) (10.a)

Re_e = {-5.57299 + 0.06035 Tm - 2.38881 10-5 Tm2} (10-4) (10.b)

k_e = {5.06021 - 0.04244 Tm + 1.96459 10-4 Tm2 - 4.19064 10-7 Tm3 (10.c)

+ 3.4922 10-10 Tm4} (10-2)

25
where Tm is the average temperature of the thermoelectric module in degrees Kelvin.

5.2. Cooling Channel

The channel cooled surface by each thermoelectric cooling module is assumed

isothermal at the temperature of the associated module. Each module is considered

independent from the other modules. The air flow in the channel is modeled as 1-D

developing flow considering that the channel width is much smaller than its length.

There will be an initial length of the channel where sensible cooling will take place until

the TE module cold surface temperature is equal to the dew point of the passing air

flow. Once dew point is reached at the surface, it is treated as wet over the rest of the

modules in the channel [22]. The pumping cooling capacity at the cold module side is

given by rate of enthalpy change given by

Qc = ̇ (ha-in – ha-out) (11)

where ̇ is the air mass flow rate ha-in and ha-out are the enthalpies of the air at inlet and

exit of each module part of the channel. The heat dissipated at the hot side is given by

(Th  Tamb )
Qh  (12)
R0

where Tamb is the ambient air temperature, R0 is overall thermal resistance on the hot

side. The variation in air temperature over the dry part of the channel for each module

sector of the channel is given by

Ta  in  Tc
 exp( 
Ahi
Ta  out  Tc
) (13.a)
m c p

The variation in the air enthalpy along each of all the wet modules is given by

26
ha  in  hsurface
 exp( 
Ahi
ha  out  hsurface
) (13.b)
m c p

where cp is the specific heat capacity in J/kgK, A is the cooled surface area, hi is the

overall internal convective heat coefficient inside the duct which is calculated and

updated in both dry and wet conditions, Ta-in and Ta-out are the temperatures of inlet and

exit air flow of the ducted module while ha-in and ha-out are the enthalpies of inlet and

exit air flow of the wet ducted module, and hsurface is the saturation enthalpy of the

flowing air at the cold surface temperature.

The workable range of air flow in the TEC channel for dehumidification to take place

corresponds to transitional and turbulent flow ranging from 5,000 to 20,000 based on

the hydraulic diameter. By design, almost half of the channel length will be dry and the

rest is wet. The temperature of the modules of the dry section is almost isothermal with

difference between first module surface temperature and last dry module temperature is

less than 1 °C. The wet section temperature difference between first and last wet

modules is less than 1.6°C. The developing turbulent flow correlation of average

Nusselt in terms of Reynolds and Greatz number over each section was used to estimate

the internal convection coefficient [22].

The internal heat convective transfer coefficient for the wet part of the cooling channel

can be obtained by the following equation

hi wet  hi dry  Cd w (14)

where hi_dry is the internal heat convective transfer coefficient for the dry part and Cd-w

is given by

Cd w  
dh 1
(15)
dT cp m

27
Hence, is the change in the enthalpy with respect to the change in temperature and it

is obtained by the following correlation based on the cold surface temperature (Tc) [23]

 3.5625 10 7  Tc 4 3.646 10 5  Tc  0.0006939  Tc  0.5472  Tc  1.667 (16)

dh 3 2

dT

Finally, the specific heat of moist air is determined by the following equation

cp m  cp  w  cp w (17)

The amount of water condensate obtained during a period of time t is given by

mCond  m  (wain  waout )  t (18)

where w is air humidity ratio in kg(water)/kg (air), cpw is the water vapor specific heat

capacity in J/kgK.

The channel flow model will be integrated with the solar still humidification

model developed by Alsaidi et al. [19]. The solar distiller air flow and brackish water

basin are modeled as lumped quasi-static systems. The solar distiller model predicts

well the exit air temperature and humidity from the distiller. It assumes a well-insulated

solar distiller with a lumped glass temperature. In addition, the solar distiller model

assumes that air is a low capacity fluid and neglects the spatial and transient variation in

the temperature of air. The model is based on a set of energy balances developed for the

distiller glass, the air flow, the water in the basin, and the water vapor flow.

The experimentally validated solar still model of Alsaidi et al. [19] predicts the

humidified air exit temperature and humidity for an open cycle. In our work, the air

input to the solar distiller model is not the ambient air, but is a mixture of fresh air from

outside and return air from the distiller. This will increase the water content of the air

leaving the distiller to the TEC channels.

28
 CHAPTER 6

NUMERICAL SOLUTION METHODOLOGY

The set of energy balance equations of the three system units of thermoelectric

cooling modules, cooling channel, and solar distiller described in the previous section

will be solved numerically to predict the exit air conditions from the cooling channel.

Input parameters to the distiller unit are time dependent ambient air temperature and

humidity and intensity of solar radiation. Given ambient and solar conditions and

mixing ratio of re-circulated air flow to ambient flow, the solar distiller model predicts

the temperature and the water content of the air exiting the solar distiller and entering

the thermoelectric cooling channel.

Every thermoelectric cooling module is characterized by restrictions on its

operational parameters which dictate the accurate selection of the optimal cooling

module for the cooling application. Given the characteristic thermoelectric cooling

module values as provided by manufacturer on Imax, Umax, Qmax, and ΔTmax at the hot side

temperature (Th), we can obtain other design parameters. These include the number of

thermocouples (Nb) in the thermoelectric cooling module, the geometry factor of the

thermoelectric element (Gf), the module width, length, thickness, and the module

overall thermal resistance. The overall thermal resistance is measured experimentally

and a statistical correlation for this resistance is developed for use in our model in terms

of the two main governing parameters, the input electric current and the ambient

environment temperature. In determining the thermoelectric modules’ parameters, each

thermoelectric cooling module is treated as a lumped system, and its parameters are

coupled with the next one, through the channel airflow where the air exiting after one

29
module is considered as inlet air input to the next channel section containing the

following module. Therefore, the thermal and electrical parameters will not be the same

for all the thermoelectric modules in the cooling channel because each one is working

under specifically different conditions.

Based on the mathematical formulation presented above, we have for each

module ten equations in ten unknowns which are Tc, Th, Qc, Qh, P, U, ha_out, α, k, and

Re. The Gauss-Seidal iterative method will be used to solve for the ten unknowns.

Given the mass flow rate of the air entering the thermoelectric cooling channel and the

TEC module input electric current, we assume a cold side temperature Tc. Using

Equation (13), the temperature and the enthalpy of the air exiting the thermoelectric

cooling module Ta-out and ha_out are obtained. The cooling capacity can then be found

from Equation (11). By equating Equations (5) and (12) of heat dissipation on the hot

side, and substituting α, k, and Re with their correlations in terms of the average

temperature of the module Tm, we can then calculate the value of the average

temperature of the module Tm. As a result, the hot side temperature Th and the new

updated values for the thermoelectric cooling module characteristics α, k, and Re can be

obtained in terms of the new calculated module average temperature.

Hence, we can calculate the total amount of heat dissipated at the hot side Qh

by Equation (12). The final step in the iterative method is calculating a new updated

value for the thermoelectric module cold side temperature Tc from Equation (4) and

check for convergence. Convergence is attained when difference in temperature is less

than 0.001 C after which we calculate the electric power consumption and the

coefficient of performance of the module. The procedure is repeated on the following

modules until exit air conditions from the channel are obtained. Finally, from the

30
properties of the humid air entering the cooling channel and that leaving the channel, we

can evaluate the total amount of condensate produced. A flow chart showing the Gauss-

Seidal iterative method is presented in Fig. 7.

Fig. 7: Iterative method flow chart

31
 Since the solar distiller model is a quasi-static and lumped model, it predicts

the exit air temperature and humidity from the distiller. In addition, while varying inlet

conditions coming from the distiller from hour to hour, the variables are considered

constant during every hour of analysis and study. The presented numerical model

predicts every hour the temperature and the humidity of the air leaving the cooling

channel, the variation of the air temperature through the channel, the thermoelectric

modules hot and cold side temperatures and their corresponding heating and cooling

capacities as well as the electric power consumption and the amount of collected water

condensate.

32
 CHAPTER 7

EXPERIMENTAL METHODOLOGY

This section presents the experimental setup used for validating the developed

thermoelectric cooling simulation model predictions as well as testing the performance

of the cooling system and condensed water output under different operational and

ambient conditions. A system prototype was built, tested and operated under different

conditions in the climatic chamber and under different electrical current supply to study

the effect of varying the experiment conditions on the system performance and the fresh

water condensate collected.

7.1. Experiment Description

An experimental station was set up in the Energy lab in the Scientific Research

Building (SRB) at AUB. A system prototype was built and tested under different

operational and ambient conditions in the climatic chamber (HTS700B) with an

advanced temperature and relative humidity level control system. Many experiments

were conducted in order to validate the predictions of the developed thermoelectric

cooling channel simulation model, taking into consideration that the solar still model

developed by Alsaidi et al. [19] was validated with an experimental setup and was

shown to predict well the exit air temperature and humidity. The experimental setup is

an important step in our work to ensure that our developed model predicts well the

amount of water condensate, the variation in air temperature along the cooling channel

and the thermal and electrical performance of the thermoelectric cooling modules on an

hourly basis.

33
 Fig. 8: Experimental system unit

Figure 8 shows a schematic of the thermoelectric cooling channel prototype.

Mainly our system prototype will be composed of a cooling channel, thermoelectric

modules for air cooling and heat rejection, heat exchanger at the hot side to enhance

heat dissipation, a humid air source, a heater, electric power supplies, an axial fan to

supply the channel with air and a water collection graduated beaker. The humid air

source replaces the solar still in providing supply air at high humidity content to the

34
thermoelectrically cooled channel, and the electric heater is used to control the channel

inlet air temperature. In addition, the electric power supply (a DC voltage source)

replaces the photovoltaic system and provides the electricity needed to power the

thermoelectric cooling modules to pump the needed cooling capacity. The variation of

the air mass flow rate to the channel is regulated using a damper. This climatic air is

mixed with the water vapor sprayed by the humidifier and heated using the electric

heater to reach the desirable input air conditions to the cooling channel. Then it will

flow along the entire cooling channel and will be cooled and dehumidified by the

thermoelectric cooling modules in order to generate water condensate. The heat emitted

at the hot side of the modules is removed using aluminum heat sinks with integrated DC

cooling fans. The water condensate produced will be collected downwards in a

graduated beaker.

A cooling channel of 120 cm-length with a 7 cm 5 cm cross sectional area

was built and twenty thermoelectric cooling modules were integrated along one side of

the channel with one cm gap between each two consecutive thermoelectric modules.

The other three sides of the cooling channel were insulated tightly with a fiber glass

sheeting in order to preserve the cooling capacity inside the duct.

7.1.1. Climatic Chamber

The air-conditioning system installed in the climatic chamber (HTS700B) will

help in providing the suitable ambient air temperature and relative humidity level for

our thermoelectric cooling system operation. As shown in the figure below, the chamber

is equipped with an advanced humidity and temperature control system (HTS700). In

the chamber, any component or appliance up to 1.5 kW added heat load, may be tested

35
in a controlled environment between 10°C and 43°C and within a humidity range of 45

to 75% RH.

Fig. 9: HTS700B climatic chamber with the HTS700 temperature and relative humidity
control system

7.1.2. Thermoelectric Cooling Modules

In our experimental setup, we will be using twenty TEC1-12714

thermoelectric cooling modules fabricated by Thermonamic Electrical Systems. The

twenty thermoelectric cooling modules will be installed along one side of the built well

insulated cooling channel. Each TEC1-12714 thermoelectric cooling module consists of

36
127 couples of (n-p) semiconductors. As shown in the figure below, it is 50 mm 50

mm area with a thickness of 4.3 mm.

Fig. 10: TEC1-12714 geometric characteristics

TEC1-12714 are single stage Bi2Te3-based cooling modules designed for both

cooling and heating applications depending on the polarity of the DC voltage supply

source. The two ceramic substrates joining the thermoelectric elements in the module is

formed of Alumina (Al2O3, white 96%). These cooling modules can be used for cooling

down to -20°C and heating up to 100°C. In addition, those cooling modules are

designed for many applications including food and beverages refrigeration services,

liquid cooling, portable cooler boxes for cars, CPU coolers and for cooling scientific

instruments and medical systems and devices.

The performance characteristics of the (n) and (p) semiconductor elements

forming the thermoelectric modules at a temperature of 300 K are given in Table 1:

37
 Table. 1: Performance specifications for the (n) and (p) semiconductor elements

Performance Specification p-type n-type

Density (g.cm-3) 6.8 7.8

Seebeck Coefficient (µV.K-1) 190 230 180 230

Electrical Conductivity (102.ohm. m-1) 850 1200 850 1200

Thermal Conductivity (W.m-1.K-1) 1.2 1.8 1.2 1.8

The overall performance specifications for the whole TEC1-12714

thermoelectric cooling module including the maximum temperature difference ΔTmax,

maximal operating current Imax , maximal applied voltage difference Umax and maximum

cooling capacity Qcmax based on the module hot side temperature Th are given in Table

2:

Table. 2: TEC1-12714 performance specifications

TEC Performance Parameters’ Specifications

Th (Hot Side Temperature in °C) 27 50

ΔTmax (Temperature Difference in °C) 68 76

Umax (Voltage in Volts) 14.8 16.6

Imax (Electric Current in Amperes) 14 14

38
 Qcmax (Cooling Capacity in Watts) 135 148

AC Resistance (ohms) 0.95 1.06

7.1.3. Heat Sinks

Twenty 55*55 mm HXHF CPU coolers formed of an aluminum heat sink and

an integrated DC fan will be equipped and placed at the hot side of the thermoelectric

modules to repel heat out. Each CPU cooler will be mounted at the hot side of a

thermoelectric module and conductive silicone-based thermal grease (silica gel) will be

filled at the module-cooler interface to ensure that no thermal gaps will occur.

Fig. 11: The 55*55 mm HXHF CPU coolers used with the heat sinks and the DC fans

39
7.1.4. Electric Power Supplies

Four electric power supplies will be used to feed the thermoelectric modules in

a manner that each power supply will be responsible to supply electrical power to five

thermoelectric modules in series with a constant DC voltage. This will allow us to

control more precisely the voltage difference applied and the electric current input to the

thermoelectric cooling modules. An additional electric power supply will be used to

feed the DC cooling fans integrated on each heat sink of the CPU coolers.

7.1.5. Heater and Humidifier

An electric humidifier will be responsible for increasing the water content of

the air flowing inside the thermoelectric cooling channel. The humidifier disperses

water vapor in the form of small droplets through the channel to mix with the ambient

input air. This amount of water vapor sprayed could be controlled manually through the

damper as shown in Fig. 8. In addition, we will be using an electric heater which is

mounted at the inlet of the thermoelectric cooling channel as shown in Fig. 8. The role

of the heater is to control the input air temperature entering the thermoelectric cooling

channel reaching the desired temperature. Therefore, by utilizing both the electric heater

and the electric humidifier, we can adopt to the needed input air conditions to

thermoelectric cooling channel by controlling the temperature and the relative humidity

of the air entering the channel to the desired input conditions.

7.1.6. Axial Air Fan

An axial 12 W AC supply fan rated at maximum flow rate of 54 cfm was used

and mounted at the top of the vertical mounted cooling channel in order to supply

40
ambient air to the channel.

7.1.7. Measurements and Recordings

Different K-type thermocouples will be installed to record the temperature of

the air through the cooling channel at six positions with an accuracy of 3°C: the

temperature of the air entering the channel, the temperature of the air exiting the

channel, and the temperature of the air at four different positions along the

thermoelectric cooling channel. In addition, we will be reading the thermoelectric cold

side and hot side temperatures. Thus, we will be recording the variation of the air

temperature through the thermoelectric cooling channel at six different heights:

 At the channel inlet

 Between the 4th and the 5th thermoelectric modules

 Between the 8th and the 9th thermoelectric modules

 Between the 12th and the 13th thermoelectric modules

 Between the 16th and the 17th thermoelectric modules

 At the channel outlet

In addition, a humidity sensor will be used to record the humidity of the air

entering the cooling channel with an accuracy of 1%and a flow meter to measure the

air flow rate in the channel by recording the velocity of the air with an accuracy of

0.01 m/s. Moreover, a heat flux meter of 0.6 s response was equipped to measure the

heat flux emitted at the hot side of the thermoelectric cooling modules. The water

condensate was collected in a beaker placed on an accurate digital scale to monitor

weight of collected water. The digital precision balance is of maximum scale reading of

3500 g and accuracy of 0.01 g. Preliminary experiments were carried out to determine

41
the overall thermal resistance at the hot side of the thermoelectric modules by recording

the heat flux released at the hot side of the thermoelectric modules as well as the

temperature. The heat flux is recorded in Volts where each 1.77 µV is equivalent to 1.0

W/m2 heat flux.

7.2. Experimental Protocol

The experiments were conducted in the first two weeks of August 2010. A

system prototype was built inside the climatic chamber situated in the Energy lab in the

scientific research building (SRB) at AUB. Four sessions of experiments were carried

out as follows:

 Session 1 aims to determine experimentally the overall thermal resistance at the

hot side of the thermoelectric modules

 Session 2 aims to study experimentally the effect of varying the mass flow rate

of the input air to the thermoelectric cooling channel

 Session 3 aims to study experimentally the effect of changing the electric current

input to the thermoelectric modules

 Session 4 aims to study experimentally the effect of changing the ambient

temperature in the climatic chamber as well as the input air temperature to the cooling

channel

Several experiments were performed for different mass flow rates of inlet air,

electric current input to the thermoelectric cooling modules, and different channel inlet

air temperature and humidity. The electric current varied between 1 A to 3.5 A, channel

air mass flow rate was varied between 0.005 to 0.012 kg/s, inlet air temperature to the

cooling channel varied between 25 °C to 34 °C, inlet air relative humidity varied

42
between 70 % to 90 %, and ambient air temperature varied between 25 °C to 34 °C. The

variation in the air temperature along the thermoelectric cooling channel was recorded,

and the amount of water condensate generated was collected and measured in intervals

of one hour during the experimentation sessions

7.3. Experimental Setup

Fig. 12: The built thermoelectric cooling channel

The experimental setup comprises the following steps:

1. First, mount the thermoelectric cooling channel built in a favorable position in

the climatic chamber.

43
 2. Control the temperature and the relative humidity level inside the climatic

chamber using the HST700 control system.

3. Then, prepare all the devices and the instruments needed to be used in the

experiment including the power supplies, the electric humidifier, the electric heater and

the all the measurement devices.

4. Fill the electric humidifier with water.

5. Using electrical wires connect the thermoelectric cooling modules and the DC

fans to the electrical power supplies.

6. Turn on the power supplies and regulate the needed voltage difference applied to

the thermoelectric cooling modules and the DC fans mounted on the heat sinks.

Fig. 13: The electric power supplies used in the experimentations

7. Turn on the four series of thermoelectric cooling modules.

44
 8. Turn on the air supply axial fan mounted at the top of the thermoelectric cooling

channel.

9. Turn on the electric humidifier and the electric heater.

10. Through the damper, control the relative humidity of the input air to the channel.

11. Record the mass flow rate of the input air using an air flow meter, the

temperature of the input air to the channel as well as the variation in the air temperature

at six positions along the thermoelectric cooling channel using thermocouples, the cold

and the hot side temperature of the thermoelectric cooling modules, the relative

humidity of the input air using a humidity sensor, and the heat released at the hot side of

the cooling modules using a flux meter.

12. Collect and measure the amount of water condensate in intervals of one hour of

system operation.

These steps were repeated during the four sessions of experimentations with some

particular modifications for each set of experiments.

7.4. Experimental Results

7.4.1. Experiment Session 1

In the first Session, we were aiming to measure experimentally the overall

thermal resistance at the hot side of the thermoelectric modules. Many experiments

were conducted and we were using a heat flux meter to measure the heat released at the

hot side of the modules. In addition, we were measuring the temperature of the hot side

of the modules as well as the ambient air temperature using K-type thermocouples. As a

result we were able to obtain the overall thermal resistance at the hot side of the

thermoelectric modules.

45
 The heat flux meter used is a self-generating device giving readings on a

microvolt basis. So, by using a multimeter to read the voltage output from the heat flux

meter, we were able to deduce the heat flux emitted in Watts based on the sensor

calibration report. Many experiments were done and it is found that the two dominating

parameters affecting the thermal resistance are the electric current input to the

thermoelectric cooling modules and the ambient air temperature The experiments were

done for the ranges of ambient and operational parameters of electric current [2-3.5 A]

and climatic chamber ambient air temperature [25-32 °C]. Table 3 presents the

experimental results obtained:

Table.3: Experimental results of the overall hot side thermal resistance

Current Tamb Reading Qhot Thot Resistance
(A) (°C) (mV) (W) (°C) (K/W)
2 25 15.2 12.881 36.1 0.861
2.5 25 21.2 17.966 40.7 0.873
3 25 23.9 20.254 43.1 0.893
3.5 25 27.3 23.135 46.2 0.916
Current Tamb Reading Qhot Thot Resistance
(A) (°C) (mV) (W) (°C) (K/W)
2 28 16.1 13.644 40.1 0.886
2.5 28 22.6 19.152 45.2 0.898
3 28 25.2 21.355 47.5 0.913
3.5 28 29.1 24.661 50.8 0.924
Current Tamb Reading Qhot Thot Resistance
(A) (°C) (mV) (W) (°C) (K/W)
2 30 17.2 14.576 43.1 0.898
2.5 30 23.8 20.169 48.3 0.907
3 30 28.5 24.152 52.3 0.923
3.5 30 33.1 28.050 56.2 0.934

46
 Current Tamb Reading Qhot Thot Resistance
(A) (°C) (mV) (W) (°C) (K/W)
2 32 18.3 15.508 46.1 0.909
2.5 32 24.5 20.762 51.1 0.919
3 32 29.8 25.254 55.5 0.930
3.5 32 34.4 29.152 59.6 0.946

Based on the experimental results presented in the table above, we were able to

develop a statistical correlation for the overall thermal resistance at the hot side in terms

of the ambient air temperature (Tamb) and the input electric current (I). This correlation

will be integrated in our developed simulation model and will be used in the model

predictions:

R = 0.6790632254 + (0.005436093 Tamb) + (0.0263714047 I) in (K/W) (19)

R2-value = 0.98

7.4.2. Experiment Session 2

In the experimental Sessions 2, 3 and 4, we will be measuring the amount of

water condensate collected and the variation in the air temperature along the

thermoelectric cooling channel at six positions or heights as follows:

 At the channel inlet designated by (T1)

 Between the 4th and the 5th thermoelectric modules designated by (T2)

 Between the 8th and the 9th thermoelectric modules designated by (T3)

 Between the 12th and the 13th thermoelectric modules designated by (T4)

 Between the 16th and the 17th thermoelectric modules designated by (T5)

 At the channel outlet designated by (T6)

47
 In Session 2, we were aiming to study experimentally the effect of varying the

electric current input to the thermoelectric cooling modules on the air temperature

variation inside the cooling channel and the water condensate output from the system.

So, we have maintained constant air mass flow rate of 0.01 kg/s, input air temperature

and relative humidity of 30°C and 85% and constant ambient air temperature of 28 °C,

while varying the input electric current to the thermoelectric modules in the range of [1-

3.5 A]. The amount of condensate was measured and the variation in the air temperature

along the channel was recorded at six different heights for six different cases. The

experimental results are presented in Table 4:

Table. 4: Water condensate and air temperature variation for different input electric
current

Electric Condensate
T1 T2 T3 T4 T5 T6
Current (A) (kg)

1 0.126 30 28.7 27.6 26.3 24.9 24.6

1.5 0.195 30 28.3 26.4 24.5 23.2 22.8

2 0.243 30 27.8 25.8 24.1 22.6 20.8

2.5 0.281 30 27.4 25.1 23 21.4 20.3

3 0.301 30 27.1 24.1 22.7 20.4 19.4

3.5 0.314 30 26.9 24.2 21.6 20.2 18.7

Figure 14 shows the increase in the amount of water condensate with the increase in the

electric current input to the thermoelectric cooling modules.

48
 0.35

0.3
Water Condensate (kg)

0.25

0.2

0.15

0.1

0.05

0
0 0.5 1 1.5 2 2.5 3 3.5 4

Electric Current (A)

Fig. 14: Amount of water condensate versus the input electric current

Figure 15 shows the variation in the air temperature along the cooling channel for

different input electric currents.

31
1A 1.5 A 2A 2.5 A 3A 3.5 A

29

27
Air Temperature (°C)

25

23

21

19

17

15
T1 T2 T3 T4 T5 T6
Temperature Recording Position
Fig. 15: Air temperature variation with electric current input

49
7.4.3. Experiment Session 3

In Session 3 of the experiments, we were aiming to study experimentally the

effect of varying the mass flow rate of the air input to the thermoelectric cooling

channel on the variation of air temperature inside the cooling channel and the water

condensate output from the system. So, we have maintained constant electric current

input to the thermoelectric modules of 2 A/module, input air temperature and relative

humidity of 30°C and 85% and constant ambient air temperature of 28 °C, while

varying the input air mass flow rate in the range of [0.005-0.011 kg/s]. The amount of

water condensate was measured and the variation in the air temperature along the

channel was recorded at six different heights for seven different cases. The experimental

results are presented in the Table 5:

Table. 5: Water condensate and air temperature variation for different input air mass
flow rate

Mass Flow Condensate

T1 T2 T3 T4 T5 T6
Rate (kg/s) (kg)

0.005 0.181 30 26.8 24.1 21.3 18.8 17.5

0.0059 0.192 30 27 24.3 21.6 20.2 18.4

0.0067 0.202 30 27.2 24.5 22.1 20.7 19.3

0.0076 0.219 30 27.4 24.6 23.2 21.3 19.6

0.0084 0.227 30 27.6 25.1 23.3 21.4 20.4

0.0092 0.237 30 27.7 25.5 23.5 22.1 21

0.0101 0.241 30 27.8 25.7 24.1 22.2 21.4

50
Figure 16 shows the increase in the amount of water condensate with the increase in the

input air mass flow rate to the thermoelectric cooling channel:

0.3

0.25

0.2
Water Condensate (kg)

0.15

0.1

0.05

0
0.004 0.005 0.006 0.007 0.008 0.009 0.01 0.011

Air Mass Flow Rate (kg/s)

Fig. 16: Amount of water condensate versus the air mass flow rate

Figure 17 shows the variation in the air temperature along the cooling channel for

different input air mass flow rates:

51
 0.005 kg/s 0.0059 kg/s 0.0067 kg/s 0.0076 kg/s

31 0.0084 kg/s 0.0092 kg/s 0.0101 kg/s

29

27
Air Temperature (°C)

25

23

21

19

17
T1 T2 T3 T4 T5 T6
Temperature Recording Position

Fig. 17: Air temperature variation with air mass flow rate

7.4.4. Experiment Session 4

In Session 4 of the experiments, we were aiming to study experimentally the

effect of changing the climatic chamber ambient air temperature and the air input

temperature to the thermoelectric cooling channel on the air temperature inside the

cooling channel and the water condensate output from the system. So, we have

maintained a constant electric current input to the thermoelectric cooling modules of 2

A/module, input air relative humidity of 80% and constant input air mass flow rate of

52
0.01 kg/s, while changing the input air temperature in the range of [28-34 °C] and the

climatic chamber ambient air temperature in the range of [26-32 °C]. The amount of

condensate was measured and the variation in the air temperature along the channel was

recorded at six different heights for seven different cases. The experimental results are

presented in Table 6:

Table. 6: Water condensate and air temperature variation for different ambient and input
air temperature
Ambient Air
Condensate
Temperature T1 T2 T3 T4 T5 T6
(kg)
(°C)

26 0.102 28 26.2 23.9 22.6 22.4 21.8

27 0.136 29 27.1 25.1 23.8 22.3 21.9

28 0.168 30 28.1 25.5 24.4 23.3 22.4

29 0.197 31 28.9 26.8 24.9 23.1 22

30 0.239 32 29.4 27.5 25.8 23.4 22.7

31 0.271 33 30.6 28.5 25.9 24.5 23.1

32 0.308 34 31.5 29.4 26.7 24.8 23.5

Figure 18 shows the increase in the amount of water condensate with the increase in the

input air temperature to the cooling channel and the ambient air temperature:

53
 0.35

0.3
Water Condensate (kg)

0.25

0.2

0.15

0.1

0.05

0
26 27 28 29 30 31 32 33 34 35 36

Input Air Temperature (°C)

Fig. 18: Amount of water condensate versus the input air temperature
The chart below shows the variation in the air temperature along the cooling channel for

different input air temperature:

54
 Ta_in = 28 Ta_in = 29 Ta_in = 30 Ta_in = 31

Ta_in = 32 Ta_in = 33 Ta_in = 34

34

32
AIr Temperature (°C)

30

28

26

24

22

20
T1 T2 T3 T4 T5 T6
Temperature Recording Position

Fig. 19: Air temperature variation with air mass flow rate

7.5. Simulation Model Validation

Figure 20 shows the amount of water condensate collected per hour as a

function of (a) electric current, (b) air mass flow rate, and (c) inlet air temperature for

experimentally measured and numerically predicted values. The simulation model

predicts well the amount of water condensate produced by the new integrated system

over hourly intervals with a maximum relative error less that 8 %. Figure 21 presents

the variation in the air temperature through the thermoelectric cooling channel recorded

at six different heights: at the channel inlet (T1), between the 4th and the 5th

thermoelectric cooling modules (T2), between the 8th and the 9th modules (T3), between

55
the 12th and the 13th modules (T4), between the 16th and the 17th modules (T5) and at the

channel outlet (T6),at different electric current input to the thermoelectric modules for

the case of 28°C ambient temperature, 0.01 kg/s air mass flow rate at 30°C inlet air

temperature. Our thermoelectric cooling channel simulation model predicted well the

variation in the air temperature through the thermoelectric cooling module with a

maximum relative error less than 2.4 % (±0.55°C).

56
Fig. 20: Plots of (a) water condensate versus electric current as predicted by the model
and recorded experimentally, (b) water condensate versus mass flow rate as
predicted by the model and recorded experimentally and (c) water condensate
versus inlet air temperature as predicted by the model and recorded
experimentally

57
Fig. 21:Air temperature variation through the cooling channel for different electric
current inputs, 0.01 kg/s air mass flow rate, 28 °C ambient temperature and 30
°C inlet air temperature (model predictions-experimental data comparison)

58
 CHAPTER 8

CASE STUDY

To assess the viability of sizing and using the proposed integrated

thermoelectric-photovoltaic system in meeting the fresh water needs in the Lebanese

humid climate, a case study of a typical residential space of 80 m2 is considered. The

system will be sized to provide at least 10 liters of fresh water over a period of operation

from 10:00 am to 7:00 pm during the summer months from June till October. An

optimization problem is formulated such that the energy consumption of the unit is

dictated by fresh water demand and the availability of electrical energy supplied by the

photovoltaic system. The integrated system design, operation and optimization are

performed for the Beirut local humid weather conditions. The system is sized such that

multiple thermoelectric cooling channels of 120 cm length and 75 cm2 cross sectional

area with 20 thermoelectric cooling modules per each cooling channel as described in

the previous section. Moreover, the thermoelectric cooling modules use solar

photovoltaic modules to provide them with the needed electrical input. Five UE 200 M5

solar photovoltaic modules are considered for the case study to provide a 1 kW capacity

photovoltaic system. The number of channels needed and their associated number of

thermoelectric cooling modules are dictated by two main constraints as follows:

 The availability of 1 kW of solar photovoltaic modules capacity.

 The amount of fresh water produced over the 10 hours of daily operation.

A solar distiller of 1.2 m2 area with a constant air mass flow rate of 0.15 kg/s is

integrated with the thermoelectric-photovoltaic system to provide desired humid air

flow to the cooling channels and enhance fresh water production. The solar distiller air

59
flow is partially re-circulated while only a fraction of this flow is drawn from ambient

air equal to the dehumidified air stream supplied to the thermoelectric cooling channels

(see Fig. 4). This modification helps in increasing the system fresh water output by

increasing further the water content in the supply air to the dehumidifying

thermoelectric channel.

The optimization problem is controlled by the air mass flow rate of the

thermoelectric channels and the electric current input to the thermoelectric cooling

modules. A minimum of five thermoelectric cooling channels are needed to produce the

10 liters of water over 10 hours of operation. Many simulations were performed, at

different air mass flow rates from 0.005 kg/s to 0.02 kg/s, to find the optimal humid air

mass flow rate to be introduced to the thermoelectric cooling channels from the solar

distiller based on the electrical energy consumption of the system and the amount of

water condensate produced. It was found that the amount of water produced increases

with the increase in the air mass flow rate until the air mass flow rate reaches 0.0155

kg/s per channel. Figure 22 shows the variation of water condensate produced with the

variation in the air mass flow rate in the thermoelectric cooling channels for different

electric current input in the month of August. For air mass flow rates higher than 0.0155

kg/s, a decrease is exhibited in the amount of fresh water produced by the system. Based

on these results, an air mass flow rate of 0.0155 kg/s will be adopted per cooling

channel.

60
 20

18

16
Water Condensate (liters)

14

12

10

0
0.010 0.011 0.012 0.013 0.014 0.015 0.016 0.017 0.018 0.019 0.020
Air mass flow rate (kg/s)

1 A Electric Current 2 A Electric Current 3 A Electric Current

Fig. 22: Variation of water condensate with the variation of air mass flow rate under
different electric current supply to the system in August

The optimal electric current input to the thermoelectric cooling modules is

selected such that a minimum of 10 liters of fresh water is obtained at the maximum

electrical energy available from the photovoltaic system. The first step in solving for the

optimal settings is to calculate the maximal electrical energy generated by the 1 kW

photovoltaic system in the Lebanese summer humid climate. This is governed mainly

by the solar irradiation intensity and the ambient air properties that vary significantly

from one month to another. As a result, this maximal energy is calculated for each of the

five sunny months from June till October. Simulations were performed while varying

the electrical current input to the thermoelectric cooling modules with a step of 0.1 A in

the range of 1 to 3 A. Based on the electrical energy consumption results and the

61
amount of fresh water collected, we can determine the optimal electrical current input to

the system for each month.

Figure 23 (a-b) shows the amount of fresh water produced and the PV power

consumption as a function of electric current input to the thermoelectric modules for (a)

August and (b) September. On the same plots, the constraint values for minimum water

production and maximum power consumption are indicated. It is clear that the month of

August provides wider operating range (shown by the dark rectangle borders in the

figure) for the system to produce more than 10 liters per day, while in September, the

operating margin for current input is very small (1.95 to 2.05 A).

(a)

62
 (b)

Fig. 23 (a-b): The amount of fresh water produced and the PV power consumption as a
function of electric current input to the TEC modules for (a) August and (b) September.

Performing the simulations over all the summer months leads to the optimal

electric current input values of 2.2 A/module for the month of June, 2.1 A/module in

July, and 2.0 A/module for August, September and October. The selected current input

values provide the needed cooling capacity, produce the desired fresh water, and meet

the constraints at acceptable thermoelectric cooling system coefficient of performance

over the 10 hours of system operation per day for the five studied months.

63
 CHAPTER 9

RESULTS AND DISCUSSION

The hourly and overall performance of the proposed PV-powered

thermoelectric cooling system integrated with solar distiller to produce at least 10 liters

of fresh water per day are presented in this section. Our main concern in the case study

was to optimize the performance of the system in order to attain our objective to collect

at least 10 liters of fresh water during the period of 10 hours of the system operation per

day, taking into account the electrical energy available from the solar photovoltaic

system in the five studied months. Table 7 presents the hourly and the total amount of

fresh water produced for the summer months showing how the system, with the

optimized air mass flow rate and input electric current to the thermoelectric cooling

modules, met its goal of producing at least 10 liters of fresh water during the 10 hours of

operation per day. Also, Fig. 24 shows the hourly amounts of fresh water condensate

collected during the five months. It is obviously shown that the maximum amount of

fresh water produced is attained during the period of the system operation between 3:00

and 6:00 pm for the five months.

64
 Table.7: Hourly and total amount of condensed water over the five months

Month

June July August September October

Electric
Current Input 2.2 2.1 2 2 2
(A)

Hour Fresh Water Production (ml)

10:00 AM 714.202 659.889 844.194 547.186 516.379

11:00 AM 914.973 796.590 951.025 581.769 727.313
12:00 PM 1028.959 1016.686 1042.092 781.067 923.350
1:00 PM 1172.997 1136.452 1303.789 958.203 1034.387
2:00 PM 1285.291 1328.650 1404.791 1025.246 1215.773
3:00 PM 1296.410 1496.495 1528.136 1292.842 1310.552
4:00 PM 1379.946 1654.413 1546.405 1345.184 1342.262
5:00 PM 1352.033 1647.909 1937.384 1308.267 1168.666
6:00 PM 1272.127 1617.546 1861.171 1255.802 1025.305
7:00 PM 1077.185 1536.688 1672.677 1113.221 803.444

Total (ml) 11494.123 12891.319 14091.664 10208.788 10067.432

65
 2

1.8

1.6
Water Condensate (liters)

1.4
june
1.2
july
1
august
0.8
september
0.6
october
0.4

0.2

0
10 11 12 13 14 15 16 17 18 19
Hour

Fig. 24: Hourly fresh water production

In addition, and based on the results obtained and shown in Fig. 25, the total

amount of fresh water produced ranged from about 14.09 liter/day in August to about

10.06 liter/day in October during 10 hours of the system operation. This is mainly due

to the humid ambient air conditions and the higher solar radiation intensity which vary

significantly from one month to another.

66
 16

14

12
Water Production (Liters)

10

0
June July August September October

Fig. 25: Daily fresh water production for the five months

Figure 26 compares the total amount of fresh water collected for the PV-

powered thermoelectric cooling system operation with and without integration with the

solar distiller. The requirement for producing the 10 liters of water cannot be met for all

the months without the humidification process through the solar distiller. The solar

distiller enhances the water condensate yield by increasing the water content of the air

introduced to the thermoelectric cooling channels. In addition, the step of re-circulating

a portion of this humid air back to the solar distiller has proven to be a very wise

decision enhancing further the water condensate produced. The amount of water

condensate produced due to the integration of the solar distiller in the process has

increased from 4.72 liters to 14.09 liters in August and from 2.31 liters to 10.06 liters in

October.

67
 16

14

12
Water Production (Liters)

10

0
June July August September October

with solar distiller without solar distiller

Fig. 26: Total fresh water production over 10 hours of the PV-powered TEC system
with and without the solar distiller

68
 10

7
Cooling Capacity (kWh)

0
June July August September October

(a)

6
Energy Consumption (kWh)

0
June July August September October

(b)

Fig. 27(a-b): Plots of (a) the cooling capacity pumped by the thermoelectric cooling
modules/ day and (b) the associated system energy consumption from the
PV in kWh/day over the summer months.

69
 Fig. 27 presents the amount of cooling capacity pumped by the thermoelectric

cooling modules situated along the thermoelectric cooling channels. This cooling

capacity reaches its maximum in June with about 9.07 kWh/day while decreases to

about 7.02 kWh in August. Furthermore, the maximum system coefficient of

performance is attained in October with a COP of 1.52. Thus, although the amount of

fresh water produced in October is the lowest among all the five months studied, we can

notice that the month of October exhibits the highest COP values, meaning the highest

cooling capacity pumped compared to the electrical consumption of the thermoelectric

cooling channels. This is based on the relatively lower ambient temperature values in

October compared to the other months, and thus these lower ambient temperature values

allow attaining lower temperature gradient between the two sides of the thermoelectric

cooling modules (the hot and the cold side). And based on the thermoelectric modules

performance characteristics, the cooling capacity pumped or the amount of heat

extracted is inversely proportional to the temperature gradient between the two

thermoelectric sides. Therefore, lower temperature gradient values in October provides

better and more appropriate performance of the thermoelectric cooling modules

allowing more heat to be extracted, and as a result more cooling capacity is pumped.

In addition, Fig. 27 presents the associated system energy consumption from

the PV in kWh/day over the summer months during 10 hours of the system operation.

The energy consumption varies from a minimum 5.32 kWh/day in October to a

maximum value of 6.73 kWh/day in June. Regarding the optimal operational energy

consumption of the system, we will compare this energy consumption with that of the

typical conventional atmospheric water extraction machines which use a vapor

compression cycle for residential applications. These conventional vapor compression

70
systems consume about 0.7 kWh to produce 1 liter of fresh water. However, the energy

intensity of our proposed integrated solar distiller with PV-powered TEC channel

system is found minimum at 0.387 kWh/liter in August and maximum in June at 0.586

kWh/liter.

9
Cost of Water Production (cents/liter)

0
June July August September October

Proposed System Conventional System

Fig. 28: The cost of fresh water production by the system for five months compared to
the cost for the conventional vapor compression systems

As shown in Fig. 28, the cost of production of one liter of fresh water was

about 7.03 cents/liter in June and boils down to 4.65 cents/liter in August at 0.12

$/kWh. On the other hand, the cost of production of one liter of fresh water by the

conventional vapor compression water making machines is in the range of 8.5

cents/liter. Therefore, comparing the costs, we can notice that the proposed system

allows significant savings on the energy consumption ranging between 17.26 % in June

71
and 45.25 % in August, compared to the vapor compression atmospheric water

extraction residential systems.

72
 CHAPTER 10

CONCLUSION AND RECOMMENDATIONS

A new integrated solar-driven thermoelectric cooling system for air

dehumidification and fresh water production was presented and experimentally

validated in this work. The system is totally powered by renewable energy sources. A

model of the integrated system is developed and is validated experimentally. The air

flow rate, air inlet conditions and electric current input to the thermoelectric cooling

modules are the controlling parameters for optimal system operation to meet desired

need of water condensate. The developed numerical simulation model predicted well

the temperature and the humidity of the air leaving the channel, the variation of the air

temperature through the whole channel, the thermoelectric modules’ hot and cold-side

temperatures and their corresponding heating and cooling capacities as well as the

electric power consumption. The thermoelectric cooling model was validated

experimentally and found to predict well the variation in the air temperature through the

thermoelectric cooling channel with a maximum relative error less than 2.4 %

(±0.55°C). In addition, the simulation model predicts well the amount of water

condensate produced by the new integrated system with a maximum relative error less

that 8 %.

A case study is presented demonstrating the feasibility of implementing the

integrated solar distiller PV powered TEC channel system in Beirut climate for

producing at least 10 liters of water per day over the summer months. The aim of the

case study was to predict the overall system performance in terms of the amount of

water condensate and energy consumption. It was found that the optimal operation total

73
energy consumption varied between 5.32 kWh/day in October and 6.73 kWh/day in

June. Moreover, the cost of production of one liter of fresh water was about 6.34

cents/liter in October with about 17.26 % saving compared to the vapor compression

atmospheric water extraction residential systems, and the cost boils down to 4.65

cents/liter in August with about 45.25 %.

The new proposed thermoelectric-photovoltaic fresh water production system

has many advantages. First of all, the proposed system is powered totally by renewable

energy sources depending solely on solar energy which is used by the solar distiller and

by the photovoltaic solar panels. Moreover, the thermoelectric cooling modules used are

compact, maintenance-free, gravity independent, durable, acoustically silent reliable and

environmentally friendly, giving more reliability and durability to the system. In

addition, the proposed system will help in solving a part of the fresh water scarcity

problem in our region in an efficient and environmental friendly manner especially for

stand-alone remote areas applications in the humid climates and in the areas where the

electrical supply is intermittent. Although the initial investment cost of the system is

relatively high, the proposed system would be a long term cost saving system because

the solar energy source is free and the system components generally need no

maintenance.

An alternative study could be done for the same system by considering a

vertical upward flow of the humidified air through the thermoelectric cooling channel

rather than the vertical downward flow presented in the study. Another alternative may

be considered through studying the effect of using a more complex PV system, with

storage batteries and advanced charge controller, on the overall performance of the

system. In addition, a possible improvement to the system operation is the possibility of

74
getting benefit of the heat dissipated at the hot side of the thermoelectric cooling

modules in heating the ambient air before introducing it through the solar distiller. This

may help in decreasing the temperature at the hot side of the thermoelectric module and

in increasing the temperature of the ambient air introduced to the solar distiller. In

addition, and as shown in the study, decreasing the thermoelectric module’s hot side

temperature will result into more favorable conditions for the module performance;

therefore, one possible improvement to the system is making use of the cooled air

exiting the channels and re-circulating this air along the hot side of the thermoelectric

cooling channel.

75
 APPENDIX I

THE SOLAR STILL MODEL

In this appendix, we will present a brief description for the solar still model

which was integrated with the thermoelectric cooling model developed. The solar still

model is a quasi-static and lumped model which predicts well the exit air temperature

and humidity from the distiller. As shown in Fig. 29, ambient air enters the distiller at a

lower point passing over the water basin and leaving the distiller at the other higher end.

Energy balances are developed for the glass, air flow, seawater in the basin and the

water vapor flow. The following assumptions are applied:

 Water and the basin have the same temperature.

 The glass temperature is lumped.

 The evaporation rate change due to salinity change is neglected.

 The solar still is assumed well insulated.

76
 Fig. 29: Solar still sketch

The transient temperature of the glass cover can be represented by a transient

lumped energy balance. In the energy balance are the effects of air convection, radiation

with water and sky and solar radiation. The transient temperature of the glass cover can

be represented by

 I t  g Ag  hr1 (Tw  Tg ) Aw  hc  g (Tg  Tair ) Ag

dTg
mg c pg

 hr 2 (Tw  Tsky ) Ag  hc a (Tg  T ) Ag

dt (A.1)

where mg is the glass cover mass, Cpg is the glass specific heat capacity, Tg is the glass

temperature, t is the time, I(t) is the incident solar irradiation on the glass surface, Tw is

the water temperature, Tair is the air temperature inside the still, and Tis the ambient air

temperature, Ag is the glass cover area, Tsky is the sky temperature, and hc-g is the

convective coefficient between glass and air flow in the still, while hc-a is the convective

77
coefficient between glass and ambient air. The term on the right hand side represents the

following energy modes:

 Energy absorbed by the glass

 Radiative heat exchange of the glass with water

 Convective heat exchange with the flowing air

 Radiative heat loss with the sky

 Convective heat loss with the ambient air

The inner and outer radiative heat transfer coefficients hr1 and hr2 are given by

hr1  Fwg (Tw  Tg )(Tw  Tg )

2 2
(A.2)

hr2  Fgs (Tg  Tsky )(Tg  Tsky )

2 2
(A.3)

where  is the Stephan Boltzman constant,  is the glass emissivity, Fwg is the view

factor between water and glass surface, and Fgs is the view factor between the glass and

sky.

Based on an energy balance between the solar energy absorbed by the water

and the energy losses by convection to the flowing air, radiation to the glass and

evaporation, the transient water basin temperature is given by

mw  c pw  I ( t )  g w Aw  hr1 (Tw  Tg ) Aw  hc w (Tw  Tair ) Aw  mwh fg

dTw
(A.4)
dt

where mw is the water mass, Cpw is the water specific heat, g is the glass transmission

coefficient, w is the water absorption coefficient, and hc-w is the convective heat

transfer coefficient between water and air.

Neglecting the spatial and transient variation in the temperature of the air (low

capacity fluid), the net energy flow (convection energy exchange with both the water

78
basin and glass cover and energy of the evaporated water) to the air stream passing

through the solar still can be written as

hc  w (Tw  Tair ) Aw  hc  g (Tg  Tair ) Ag  mair c p Tair1  mwh fg

 mair w1c pv Tair1  mair c p Tair 2  mair w2c pvTair 2  0
air 1
(A.5)
air 2

where Aw is the water basin surface area, Cpv is water vapor specific heat, Tair1 and Tair2

are the inlet and exit air flow temperatures, hfg is the water heat of evaporation, m air is

 w is the water evaporation rate.

the air flow rate through the distiller, and m

The rate of water evaporation can be obtained from the correlation developed

Al-Shammiri et al. [24] that takes into account the water salinity

mw  (169.02  V 1.478  X 0.103P0.654 ) Aw

(A.6)

where V is the velocity of flowing air, X is the salinity concentration and P is the water

vapor difference between the flowing air and the water basin vapor.

79
 APPENDIX II

HOURLY FRESH WATER PRODUCTION DATA

In this appendix, the hourly fresh water production for the integrated

thermoelectric-photovoltaic system is presented over the five months studied.

1.6

1.4
Water Production (liters/h)

1.2

0.8

0.6

0.4

0.2

0
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 30: Hourly fresh water production in June

80
 1.8

1.6
Water Production (liters/h)
1.4

1.2

0.8

0.6

0.4

0.2

0
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 31: Hourly fresh water production in July

1.8
Water Production (liters/h)

1.6

1.4

1.2

0.8

0.6

0.4

0.2

0
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 32: Hourly fresh water production in August

81
 1.6

1.4
Water Production (liters/h)
1.2

0.8

0.6

0.4

0.2

0
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 33: Hourly fresh water production in September

1.6

1.4
Water Production (liters/h)

1.2

0.8

0.6

0.4

0.2

0
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 34: Hourly fresh water production in October

82
 APPENDIX III

HOURLY COOLING CAPACITY DATA

In this appendix, the hourly cooling capacity pumped at the cold side of the

thermoelectric cooling modules equipped in the integrated thermoelectric-photovoltaic

system is presented over the five months studied.

1000

950
Pumped Cooling Capacity (Wh)

900

850

800

750

700
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 35: Hourly cooling capacity pumped by the TEC modules in June

83
 830
Pumped Cooling Capacity (Wh)
780

730

680

630

580
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 36: Hourly cooling capacity pumped by the TEC modules in July

750
Pumped Cooling Capacity (Wh)

700

650

600

550

500
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 37: Hourly cooling capacity pumped by the TEC modules in August

84
 820

780
Pumped Cooling Capacity (Wh)
740

700

660

620

580

540

500

460
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 38: Hourly cooling capacity pumped by the TEC modules in September

900

860
Pumped Cooling Capacity (Wh)

820

780

740

700
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 39: Hourly cooling capacity pumped by the TEC modules in October

85
 APPENDIX IV

HOURLY HEAT DISSIPATION DATA

In this appendix, the hourly heat dissipated at the hot side of the thermoelectric

cooling system is presented over the five months studied.

1650

1600
Heat Disssipated (Wh)

1550

1500

1450

1400
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 40: Hourly heat dissipated by the TE cooling system in June

86
 1450

1400
Heat Disssipated (Wh)

1350

1300

1250

1200
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 41: Hourly heat dissipated by the TE cooling system in July

1300

1250
Heat Disssipated (Wh)

1200

1150

1100

1050
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 42: Hourly heat dissipated by the TE cooling system in August

87
 1350

1300

1250
Heat Disssipated (Wh)

1200

1150

1100

1050

1000
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 43: Hourly heat dissipated by the TE cooling system in September

1410

1360
Heat Disssipated (Wh)

1310

1260

1210

1160
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 44: Hourly heat dissipated by the TE cooling system in October

88
 APPENDIX V

HOURLY ENERGY CONSUMPTION DATA

In this appendix, the hourly associated thermoelectric cooling system electrical

energy consumption from the solar photovoltaic panels is presented over the five

months studied.

680

675

670
Power Consumption (Wh)

665

660

655

650

645

640
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 45: Hourly system energy consumption in June

89
 620

610

600
Power Consumption (Wh)

590

580

570

560

550
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 46: Hourly system energy consumption in July

560

550

540
Power Consumption (Wh)

530

520

510

500

490
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 47: Hourly system energy consumption in August

90
 560

550

540
Power Consumption (Wh)

530

520

510

500

490

480
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 48: Hourly system energy consumption in September

540

530
Power Consumption (Wh)

520

510

500

490

480
10 11 12 13 14 15 16 17 18 19
Time of the Day

Fig. 49: Hourly system energy consumption in October

91
 APPENDIX VI

TE-PV SYSTEM DATASHEETS

A) TEC1-12714 Thermoelectric Cooling Module

In this appendix, the TEC1-12714 thermoelectric cooling module performance

curves at different thermoelectric module’s hot side temperature as provided in the

manufacturer’s datasheets.

1- 27°C Hot Side Temperature

Fig. 50: Qc vs Voltage under various ΔT (Th = 27°C)

92
 Fig. 51: Qc vs ΔT under various currents (Th = 27°C)

Fig. 52: Voltage vs ΔT under various currents (Th = 27°C)

93
2- 50°C Hot Side Temperature

Fig. 53: Qc vs Voltage under various ΔT (Th = 50°C)

Fig. 54: Qc vs ΔT under various currents (Th = 50°C)

94
 Fig. 55: Voltage vs ΔT under various currents (Th = 50°C)

B) UE 200 M5 Solar Photovoltaic Module

In this appendix, the UE 200 M5 solar photovoltaic module electrical and

mechanical specifications are presented as provided in the manufacturer’s datasheets.

UE 200 M5 photovoltaic module, manufactured by Universal Energy, contains 60,

156×156 mmmonocrystalline solar cells. It is designed for residential, industrial and

commercial grid connected systems with efficiencies up to 14%.

95
 Fig. 56: UE 200 M5 electrical specifications

Fig. 57: UE 200 M5 mechanical specifications

96
 REFERENCES

[1] Miller JE, Review of Water Resources and Desalination Technologies, Sandia
National Laboratories, Albuquerque Sand 2003-0800, 2003.

[2] Mezher T, Fath H, Abbas Z, Khaled A. Techno-economic assessment and

environmental impacts of desalination technologies. Desalination 2011; 266
(1-3): 263–273.

[3] Rowe DM. CRC Handbook of Thermoelectric, CRC Press Inc. Boca Raton
U.S.A. Chapter 55: Modeling of thermoelectric cooling systems, 677-683;
1995.

[4] Huang BJ, Chin CJ, Duang CL. A design method for thermoelectric cooler.
International Journal of Refrigeration 2000; 23: 208-218.

[5] Khattab NM, El Shenawy ET. Optimal operation of thermoelectric cooler driven
by solar thermoelectric generator. Energy Conversion and Management 2006;
47: 407-426.

[6] Astrain D., Vian JG, Dominguez M. Increase of COP in the thermoelectric
refrigeration by optimization of heat dissipation. Applied Thermal Engineering
2003; 23: 2183-2200.

[7] Nolwenn LP, Matthieu C., Luo L, Fraisse G. Coupling of thermoelectric modules
with a photovoltaic panel for air pre-heating and pre-cooling application; an
annual simulation. International Journal of Energy Research 2008; 32: 1316-
1328.

[8] Charoenporn L., Wiset L. Atthajariyakul S. Evaluation of the thermal comfort of

a thermoelectric ceiling cooling panel (TE-CCP) System. Journal of
Electronic Materials. 2009; 38 (7):1472-1477.

[9] Tan F.L. and S.C Fok (2008). "Methodology on sizing and selecting
thermoelectric cooler from different TEC manufacturers in cooling system
design." Energy Conversion and Management". 49, 2008: 1715-1723.
[10] Xi Hongxia, Lingai Luo and Gilles Fraisse (2007). "Development and
applications of solar-based thermoelectric technologies." Renewable and
Sustainable Energy Reviews. 11, 2007: 923-936.

[11] Chang Yu-Wei, Cheng CH, Wu WF, Chen SL. An experimental investigation of
thermoelectric air-cooling module. World Academy of Science, Engineering
and Technology 2007; 33: 128-133.

97
[12] Cosnier M, Fraisse G, Luo L. An experimental and numerical study of a
thermoelectric air-cooling and air-heating system. International Journal of
Refrigeration 2008; 31: 1051-1062.

[13] Lertsatitthanakorn Charoenporn, lamul Wiset ans Surat Atthajariyakul.

"Evaluation of the thermal comfort of a thermoelectric ceiling cooling panel
(TE-CCP) System." Journal of Electronic Materials. 38, 2007 No.7.
[14] Gillot M, Jiang L. Riffat S. An investigation of thermoelectric cooling devices
for small-scale space conditioning applications in building. International
Journal of Energy Research 2010; 34 (9): 776-786.

[15] Le Pierre Nolwenn, Matthieu Cosnier, Lingai Luo and Gilles Fraisse (2008). "
Coupling of thermoelectric modules with a photovoltaic panel for air pre-
heating and pre-cooling application; an annual simulation." International
Journal of Energy Research. 32, 2008: 1316-1328.
[16] Dai YJ, Wang RZ, Ni L. Experimental investigation and analysis on a
thermoelectric refrigerator driven by solar cells. Solar Energy Materials and
Solar Cells 2003; 77(4): 377-391.

[17] Abdul-Wahab Sabah A., A. Elkamel, A. Al-Damkhi, I. Al-Habsi, H. Al-

Rubai’ey, A. Al-Battashi, A. Al-Tamimi, K. Al-Mamari, M. Chutani (2009).
"Design and experimental investigation of portable solar thermoelectric
refrigerator." Renewable Energy. 34, 2009: 30-34.
[18] Vian JG, Astrain D, Dominguez M. Numerical modeling and a design of a
thermoelectric dehumidifier. Applied Thermal Engineering 2002; 22(4): 407-
422.

[19] Alsaidi A., Ghaddar N, Ghali K. Experimental and theoretical study of an

optimized integrated solar desalination and air-conditioning unit. International
Journal of Green Energy 2010; in press.

[20] Tsai HL, Lin JM. Model building and simulation of thermoelectric module using
Matlab/Simulink. Journal of Electronic Materials 2010; 39 (9): 2105-2111.

[21] Yang B, Ahuga H, Tran N. Thermoelectric Technology Assessment Application

to Air Conditioning and Refrigeration. HVAC&R Research 2007; 14(5):.635-
653.

[22] Lienhard V John H., A Heat Transfer Textbook, Third Edition, Phlogiston Press,
New York, 2003.

[23] Kuehn HT, Ramsey JW, Threlkeld JL. Thermal Environmental Engineering. 3rd
Edition, Prentice-hall Inc. New York, 1998, Ch. 7.

[24] AI-Shammiri M, Evaporation rate as a function of water salinity, Desalination,

150 (2002) 189-203.

98