EWICON

The Electrostatic Wind Energy
Converter
electrical performance of a high voltage prototype
ii
The Electrostatic Wind Energy
Converter
Proefschrift
ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft
op gezag van de Rector Magnificus prof.dr.ir. J.T. Fokkema
voorzitter van het College voor Promoties,
in het openbaar te verdedigen,
op woensdag 10 december 2008 om 10:00 uur
door
Dhiradj DJAIRAM
natuurkundig ingenieur
geboren te Groningen
Dit proefschrift is goedgekeurd door de promotor:
Prof.dr. J.J. Smit
Samenstelling promotiecommissie:
Rector Magnificus Voorzitter

Prof.dr. J.J. Smit Technische Universiteit Delft, promotor
Prof.dr.eng. J.A. Ferreira Technische Universiteit Delft
Prof.dr. G.J.W. van Bussel Technische Universiteit Delft
Prof.dr.ir. G.P.A Bot Wageningen Universiteit en Researchcentrum
Prof.dr. W. Balachandran Brunel University, West London, United Kingdom
Dr.ir. P.H.F. Morshuis Technische Universiteit Delft
Dr.ir. J.C.M. Marijnissen Technische Universiteit Delft / University of Florida
This research project has been supported with a grant from the Dutch Programme EET
(Economy, Ecology and Technology, EETK02005)
ISBN: 978-90-8559-482-6
Copyright © 2008 by D.Djairam
Cover: An artist’s impression of an EWICON system at sea.

Design: Tomas Pasma
iv
Calvin: “I’ve been thinking, Hobbes…”
Hobbes: “On a weekend?”
Calvin: “Well, it wasn’t on purpose….”
- Calvin and Hobbes
Aan mijn ouders

mijn broertje
en mijn zusje
vi
Summary
The Electrostatic Wind Energy Converter
The increasing global demand for energy has rekindled the interest for
various forms of renewable energy production, including wind energy. In addition
to the conventional wind turbines, research efforts have been directed to other more
experimental methods of extracting energy from the wind. One of these methods is
the electrostatic wind energy converter (EWICON) in which wind energy is
converted to electrical energy by letting the wind move charged particles against
the direction of an electric field.
In chapter 1, the current global energy situation and the wind energy
situation in particular are described. Conventional and experimental forms of wind
energy extracting methods are described together with their respective advantages
and disadvantages. Subsequently, the concept of the EWICON principle is
explained and two possible implementations are discussed of which one is chosen.
Finally, a comparison is made between wind turbines and the EWICON system,
which indicates that there are significant advantages of a fully operational
EWICON system over wind turbines.
In chapter 2, the theoretical framework that governs the operation of the

EWICON system is laid out. The influence of the charge and particle size together
with the electric field strength and the wind speed on the output power of the
EWICON system is discussed. Issues such as life times of charged particles are
also addressed. An analytical model that describes the trajectories of the charged
particles is used to calculate the output power. Finally, the EWICON Performance
Index (EPI) is defined in order to rank the efficiency of different versions of the
EWICON system.
In chapter 3, a number of spraying methods to create charged particles are

discussed. All these discussed methods require liquids to create charged droplets.
Based on a number of requirements which the ideal method should adhere to, two
methods are chosen for further investigation. These methods are
electrohydrodynamic atomisation (EHDA) and high pressure monodisperse
spraying (HPMS). The spraying liquids are demineralised (saline) water, ethanol
and mixtures of water and ethanol. Using single nozzle versions of both spraying
vii
methods, droplets can be sprayed and charged and, therefore, both methods will be
used in the final versions of the EWICON system.
In chapter 4, the design of an EWICON test system is discussed, which
include insulation issues, the design and placement of electrodes and the liquid
supply system. Scaled up multi-nozzle versions of the two spraying methods are
constructed and tested with a generator that provides laminar air flow. These
experiments confirm that the wind can move the charged droplets against the
electric field while the EWICON system is still connected to earth without
electrical load. These experiments also show that currents associated with the
production of charged droplets are not always proportional to the number of
nozzles.
In chapter 5, various versions of the EWICON system are tested while

allowing the charged droplets to charge the system up to a certain potential. These
experiments were either conducted with or without an electrical load. Parameters
such as wind speed, spraying liquid and spraying nozzles have been varied. The
experimental results are statistically analysed and used in circuit equivalent models
to verify the operation of the EWICON system. In most cases, the maximum output
power was higher than the required operating power. Compared to the maximum
convertible power available in the wind, the net output power is in the order of a
few percent.
Based on these results, some considerations are given to a scaled up
version, 100 kW, of the EWCION system.
In chapter 6, this thesis concludes for both spraying methods that while a
net gain has been achieved in terms of output power, significant research must be
undertaken to make the EWICON concept commercially viable.
A number of recommendations are given to increase the output power of
the EWICON. First of all, the maximum current output of a single nozzle needs to
be increased significantly. Secondly, the proportionality between the current and
the number of spraying nozzles needs to be improved. Thirdly, spraying with the
individual nozzles needs to be optimised such that sea and tap water can be used.
Lastly, the design of the spraying systems should be such that the nozzles can be
packed more closely together.
viii
Samenvatting
Door de wereldwijde toenemende vraag naar energie is het onderzoek naar

verschillende vormen van duurzame energie weer in de volle belangstelling
gekomen, waaronder windenergie. Naast de verbetering van bestaande
windturbines vindt er ook onderzoek plaats naar nieuwe methoden voor het
benutten van windenergie. Eén van deze methoden wordt toegepast in een
elektrostatische windenergie converter (EWICON), waarin energie uit wind
gewonnen wordt door de wind elektrisch geladen deeltjes tegen de richting van een
elektrisch veld te laten bewegen.
In hoofdstuk 1 wordt de huidige mondiale energiesituatie uiteengezet

waarin met name aandacht wordt besteed aan windenergie. De voor- en nadelen
van de huidige windturbines alsmede die van andere experimentele
windenergietechnieken worden besproken. Vervolgens wordt het concept van de
EWICON uitgelegd waarbij twee mogelijke implementaties worden behandeld en
waaruit er één gekozen wordt. Uiteindelijk wordt er een vergelijking gemaakt
tussen de conventionele windturbines en de EWICON-methode, waaruit volgt dat
er significante voordelen ten opzichte van windturbines kunnen zijn indien een
volledig operationeel EWICON-systeem gerealiseerd kan worden.
In hoofdstuk 2 wordt het theoretisch fundament van de werking van het

EWICON-systeem gelegd. Het effect op het uitgangsvermogen van de EWICON
als functie van de elektrische lading en de afmetingen van de deeltjes in samenhang
met de windsnelheid en de grootte van het elektrische veld wordt behandeld.
Hierbij worden zaken zoals o.a. de levensduur van deeltjes ook in acht genomen.
Een analytisch model wordt opgesteld waarmee de afgelegde weg van de geladen
deeltjes en het uitgangsvermogen van de EWICON kunnen worden bepaald. Als
laatste wordt de EWICON Performance Index gedefinieerd, waarmee verschillende
varianten van het EWICON-systeem met elkaar vergeleken kunnen worden.
In hoofdstuk 3 wordt een aantal verschillende methoden behandeld

waarmee geladen deeltjes gecreëerd kunnen worden. Al deze methoden zijn
gebaseerd op het versproeien van vloeistof. Aan de hand van een lijst van eisen zijn
ix
twee methoden gekozen waar nader onderzoek aan verricht wordt. Deze methoden
zijn electrohydrodynamic atomisation (EHDA) en high pressure monodisperse
spraying (HPMS). De gebruikte sproeivloeistoffen zijn gedemineraliseerd (evt.
licht zout) water, ethanol en mengsels van beide vloeistoffen. Met beide
sproeimethoden blijken geladen druppels gecreëerd te kunnen worden in
testopstellingen waarbij gebruikt wordt gemaakt van één enkel vernevelaarsysteem.
Daarom zullen beide sproeimethoden gebruikt worden in de uiteindelijke versies
van het EWICON-systeem.
In hoofdstuk 4 wordt het ontwerp van het algehele EWICON-testsysteem

besproken, waarbij zaken als de isolatie, het ontwerp en de plaatsing van elektrodes
en vloeistoftoevoersysteem aan de orde komen. Systemen met meerdere
vernevelaars zijn getest in een omgeving met een laminaire windstroming. Deze
experimenten, die uitgevoerd zijn met een geaard EWICON systeem zonder
elektrische belasting, laten zien dat de wind in staat is de geladen druppels mee te
bewegen tegen de richting van het elektrische veld. Deze experimenten laten ook
zien dat de elektrische stroom, die gepaard gaat met de creatie van geladen
druppels, niet altijd evenredig is met het aantal vernevelaars
In hoofdstuk 5 worden verschillen versies van het EWICON-systeem

getest, waarbij nu het systeem elektrisch zwevend is en dus opgeladen kan worden
tot een bepaalde eindspanning door de geladen druppels. Deze eindspanning is o.a.
afhankelijk van de windsnelheid, de opgewekte stroom en de isolatie van de het
systeem. Deze experimenten zijn uitgevoerd met en zonder elektrische belasting.
Parameters zoals de windsnelheid, de sproeivloeistof en het aantal vernevelaars
zijn hierbij gevarieerd. De experimentele resultaten zijn statistisch geanalyseerd en
verwerkt in een circuit equivalent model om de werking van het EWICON-systeem
te verifiëren. In meeste gevallen was het opgewerkte vermogen groter dan het
benodigd vermogen. Vergeleken met het beschikbare vermogen in de wind, ligt het
netto uitgangsvermogen van het EWICON-systeem in de orde van enkele
procenten.
Op basis van de extrapolatie van de resultaten is er nog gekeken naar de
aspecten die een rol zouden spelen bij een opgeschaald 100 kW EWICON-systeem.
In hoofdstuk 6 wordt dit proefschrift afgesloten met de conclusie dat,

hoewel er in veel gevallen netto vermogen wordt opgewekt, er nog een veel
onderzoek uitgevoerd moet worden alvorens een EWICON commercieel haalbaar
zal zijn. Dit geldt voor beide sproeimethoden.
Om dit te bewerkstelligen wordt er een aantal aanbevelingen gedaan om
het uitgangsvermogen van het EWICON-systeem te verhogen. Als eerste moet de
opbrengst per vernevelaar verhoogd worden. Ten tweede moet de evenredigheid
tussen het aantal vernevelaars en de geproduceerde elektrische stroom verbeterd
worden. Ten derde moeten de vernevelaars zodanig aangepast worden dat het
x
succesvol sproeien van kraan- en zoutwater mogelijk is. Als laatste moeten het
ontwerp van de sproeisystemen zodanig verbeterd worden dat een hoge
concentratie van vernevelaars per oppervlakte mogelijk is.
xi
Table of Contents
Summary ................................................................................................................ vii
Samenvatting ........................................................................................................... ix
1. Introduction .......................................................................................................... 1
1.1 Energy production in general .................................................................. 1
1.2 Alternative energy sources...................................................................... 3
1.3 Wind energy in general ........................................................................... 4
1.3.1 Conventional methods.................................................................... 6
1.3.2 Drawbacks of wind turbines........................................................... 7
1.3.3 Alternative wind energy methods................................................... 8
1.3.4 Use of wind energy......................................................................... 8
1.4 A new concept, a system with very little mechanical movement ........... 8
1.4.1 Principle: work on charge carriers.................................................. 9
1.4.2 Implementation of the EWICON method..................................... 10
1.4.3 Expected advantages and disadvantages ...................................... 14
1.4.4 Goals EWICON research project ................................................. 16
1.4.5 Contents and approach ................................................................. 16
1.5 Acknowledgment .................................................................................. 17
2. Theoretical basis and modelling ......................................................................... 19

2.1 Energy and power contained in the wind .............................................. 20
2.2 EWICON model.................................................................................... 21
2.2.1 Rayleigh limit and droplet parameters ......................................... 21
2.2.2 Analytical model for simple configuration................................... 27
2.2.3 Constraints on size and charge, electrical mobility ...................... 31
2.2.4 Computation & numerical values................................................. 37
2.3 Performance Index of the EWICON system ......................................... 40
2.4 Conclusions........................................................................................... 41
3. Charged droplet creation methods...................................................................... 43

3.1 EWICON requirements on droplet creation.......................................... 44
3.1.1 Low energy consumption with respect to the energy in the wind 44
3.1.2 Controllable charge, high current output...................................... 45
3.1.3 Monodispersity............................................................................. 45
xiii
3.1.4 Operable under windy condition .................................................. 47

3.1.5 Environmentally friendly spraying liquid .................................... 47
3.2 Previously considered creation and charging methods ......................... 48
3.2.1 Droplet creation - Ultrasonic atomisation .................................... 48
3.2.2 Droplet creation - Centrifugal or rotating disk atomisation ......... 48
3.2.3 Droplet charging - Corona charging............................................. 49
3.3 Electrohydrodynamic atomisation ........................................................ 50
3.3.1 EHDA principle............................................................................ 50
3.3.2 Analytical model: EHDA ............................................................. 52
3.3.3 EHDA implementation and experiments...................................... 55
3.3.4 Conclusions on EHDA based spraying systems........................... 66
3.4 High pressure monodisperse spraying .................................................. 66
3.4.1 HPMS principle............................................................................ 67
3.4.2 Analytical model: HPMS ............................................................. 68
3.4.3 HPMS implementation and experiments...................................... 70
3.4.4 Conclusions on HPMS based spraying systems ........................... 72
3.5 Suitable charging method for the EWICON ......................................... 72
3.5.1 Summarising conclusions............................................................. 72
3.5.2 Decision........................................................................................ 73
4. EWICON system design..................................................................................... 75

4.1 General design....................................................................................... 76
4.1.1 Platform and isolation .................................................................. 77
4.1.2 Liquid supply system.................................................................... 80
4.1.3 Electrical system........................................................................... 81
4.2 Charged droplet transportation.............................................................. 81
4.2.1 Wind generator ............................................................................. 82
4.2.2 EHDA........................................................................................... 83
4.2.3 HPMS ........................................................................................... 85
4.2.4 Conclusions: charged droplet transportation of both methods ..... 85
4.3 Increased charged droplet production ................................................... 86
4.3.1 EHDA: Multi-needle systems....................................................... 86
4.3.2 EHDA: Self-adjusting nozzles wire spraying system................... 93
4.3.3 HPMS: Multi-device systems....................................................... 97
4.4 Lowering required wind speed by field grading ................................... 97
4.4.1 Analysis of droplet trajectories..................................................... 97
4.4.2 Steering electrode(s)................................................................... 101
4.4.3 Conclusions on field grading...................................................... 103
4.5 Rising EWICON potential and droplet movement ............................. 104
4.6 Overall conclusions and final experimental set-up ............................. 105
4.6.1 Conclusions ................................................................................ 105
4.6.2 Final experimental test set-up..................................................... 106
xiv
5. Testing of the complete EWICON system ....................................................... 109
5.1 Experimental method .......................................................................... 110
5.2 EWICON with single nozzle EHDA spraying systems ...................... 111
5.2.1 Single ring electrode................................................................... 111
5.2.2 Single cylindrical electrode ........................................................ 113
5.3 EWICON with multiple nozzle EHDA spraying systems .................. 114
5.3.1 Multiple needles with rod or rounded ring electrodes................ 114
5.3.2 Multiple needle with cylindrical electrode configurations ......... 117
5.3.3 Self-adjusting nozzles configuration .......................................... 119
5.4 EWICON with HPMS system............................................................. 120
5.4.1 Single spraying device................................................................ 120
5.4.2 Multiple spraying devices .......................................................... 121
5.5 Analysis of the experimental results ................................................... 121
5.5.1 EWICON circuit equivalent ....................................................... 122
5.5.2 Charging of the EWICON, output power................................... 124
5.5.3 Analysis of the charging current................................................. 126
5.5.4 Analysis of the output power...................................................... 126
5.5.5 The output power vs. the wind speed ......................................... 127
5.5.6 Efficiencies of the spraying systems, EPI .................................. 129
5.6 Conclusions......................................................................................... 131
5.7 Scaling up the EWICON..................................................................... 131
5.7.1 General considerations ............................................................... 131
5.7.2 Extrapolation based on the conducted experiments ................... 132
6. Conclusions and recommendations .................................................................. 133

6.1 Conclusions......................................................................................... 133
6.1.1 Conclusions per chapter ............................................................. 133
6.1.2 Overall conclusions on the EWICON concept ........................... 135
6.2 Recommendations............................................................................... 136
Appendix A. Betz’ law ................................................................................ 139
Appendix B. Evaporation and relaxation times of droplets......................... 143
Appendix C. Numerical modelling in MATLAB........................................ 147
Appendix D. Statistical analysis in MATLAB ............................................ 157

D.1 Rounded ring electrode – single nozzle .................................................... 157
D.2 Rounded ring electrodes – six nozzles ...................................................... 158
D.3 Output power vs. wind speed .................................................................... 159
Appendix E. LORENTZ ............................................................................. 161
xv
List of symbols and abbreviations ........................................................................ 163
List of references .................................................................................................. 165
Acknowledgments ................................................................................................ 169
Curriculum Vitae .................................................................................................. 173
xvi
Chapter 1 - Introduction
1
Introduction
“The answer, my friend, is blowin’ in the wind.”
– Bob Dylan
Wind energy has had a profound impact on the way our world has taken
form, from the first sailing ships that discovered new worlds to the wind mills that
were used to create dry lands in the Netherlands. With every new era, an inventive
way was found to make use of the energy contained in wind. In this thesis, we will
introduce an alternative method of extracting this energy.
While wind energy is often viewed as an energy source on its own,

technically speaking, wind energy is mainly a form of solar energy, because air
flow is generated due to the uneven heating of the Earth’s surface by the sun.
About 1% of the solar energy reaching the earth is transformed into wind energy.
Due to the uneven heating of the Earth’s surface, there are some locations that are
more suitable for the exploitation of wind energy, like at sea or oceans, on wide
open plains or along coastal lines.
In this chapter, firstly, we will start with a general outline of the energy
production in the world. After a quick review of the conventional methods that are
currently employed to utilise wind power, the theoretical concepts of the new
method will be explained and, finally, we will conclude by stating the goals of this
PhD research project.
1.1 Energy production in general
Worldwide energy consumption is growing and as countries like China and

India are rapidly industrialising towards western standards, the demand for energy
will become even higher. Currently, as we can see in Figure 1.1, the world primary
energy demand at this moment is roughly 150000 TWh and according to the
International Energy Agency, this demand will increase to almost 200000 TWh in
2030, see [1].
1
World Primary Energy Demand
200000
180000
Total
160000
140000 Oil
120000 Coal
TWh
100000
Gas
80000
60000 Hydro, biomass and
40000 other renewables
Nuclear
20000
0
1980 1990 2000 2010 2020 2030
Figure 1.1. World primary energy demand of the last decades and the forecast until
2030 according to the International Energy Agency, see [1].
Of the current global energy production 86 percent is delivered using fossil

based energy sources, i.e. oil, coal and gas. Another major form of energy
production is nuclear energy which, at this point in time, is of the fission variety,
which currently contributes 6 percent of the global energy production.
Both forms of energy production have their own drawbacks considered

from various points of view, which are not always technical or environmental
viewpoints. For example, with regards to fossil based energy sources, the main
issue is the finiteness of the supply of oil, gas and (to a lesser degree) coal reserves.
The most pessimistic estimation according to the International Energy Agency is
that there will be enough to sustain this consumption rate approximately 164 years
for coal, 64 for natural gas and 42 for crude oil, see [2]. Since oil and gas also serve
other purposes than the production of electricity, such as the manufacturing of
plastics, alternative means of energy production would mean a reduction of the
pressure on these fossil reserves.
Furthermore, with the burning of fossil fuels CO2 is produced. While the
full effects of CO2 production on the climatologic changes have yet to be
understood, it is internationally acknowledged that the reduction of the release of
CO2 into the atmosphere is necessary.
2
Similar considerations can be made concerning fission based nuclear

energy. Firstly, just like with the fossil based energy sources, there is a limit to the
supply of uranium and plutonium, a limit we might see most pessimistically around
2030, see [3]. Furthermore, even though many countries, like e.g. France and the
USA, utilise nuclear energy, public perception is still negative, mostly because of
the operating risks and the problems surrounding the disposal of the nuclear waste.
In general, whenever new plans for building a nuclear reactor are disclosed,
popular resistance to these plans can be expected. In addition, certain governments,
like that of Germany, have decided to shut down all nuclear reactors by 2020, see
[4].
Nuclear energy from fusion has a more positive image than its fission
counterpart, but it is still in an experimental phase and it is generally thought that it
will take at least a few decades before this technology is commercially available,
see [5].
1.2 Alternative energy sources
Considering the issues with the aforementioned sources of energy,

alternative means of energy production have been investigated. About 17% of the
global electricity production is generated through the use of alternative energy
sources, see [6]. Some examples of methods employing renewable energy sources
are:
- hydrogen/ethanol fuel cells,

- biomass energy
- photovoltaic solar cells
- hydro energy
- tidal energy
of which their share in the total world renewable energy usage, according to the
Renewable Energy Network Policy Network [7], can be seen in Figure 1.2.
3
World Renewable Energy 2006

0%
1%
Large hydropower (78.57%)
1% 0%
5%
Small hydropower (7.45%)
0%
8%
Wind turbines (7.55%)
7%
Biomass power (4.59%)
Geothermal power (0.97%)
Solar PV, grid-connected (0.52%)
Solar PV, off-grid (0.28%)

78%
Concentrating solar thermal power (0,04%)
Ocean (tidal) power (0,03%)
Figure 1.2. The world renewable energy sources broken down in categories for 2006,
according the REN21, see [7].
Most of these alternative methods of energy production are still only used
on a small scale basis. The independent economic viability of these alternative
renewable energy sources is low to non-existent and, therefore, most of these
methods need to be promoted through the use of subsidies and tax incentives.
Slowly, energy companies are trying to adopt a “greener” image and have started to
offer electricity generated through sustainable means. The rates for this “green”
electricity are usually equivalent to the rates for electricity generated through
conventional means, again promoted by tax breaks.
1.3 Wind energy in general
Firstly, we will briefly mention the current status of the use of wind
energy, its current capabilities, its drawbacks and what can be expected of wind
energy. Then, we will introduce a different concept of converting wind energy to
electrical energy in order to address some of the drawbacks currently associated
with wind energy production.
Usually, when it comes to alternative means of generating energy, several

terms are used to label these alternative means using adjectives like “sustainable”
or “renewable” and sometimes these adjectives are used indiscriminately.
4
Therefore, for the sake of clarity, the following definitions, which are commonly
accepted, will be used throughout this thesis:
Sustainable energy: A form of energy generation which can be

maintained economically without depleting or
damaging the resources.
Renewable energy: A form of energy generation in which the required

resources are normally replenished through natural
processes.
Usually, this means that forms of renewable energy are also sustainable if
they can also be maintained economically.
Of all the various forms of sustainable energy generation, wind energy is
one of the most utilised forms, together with hydro energy. It is a growing source
of sustainable energy which has the potential to ease the pressure on fossil based
energy sources. The fact that there is virtually no CO2 emission when generating
wind energy also means that this form of power generation could play an important
role in global energy supply especially considering the issue of global warming.
Figure 1.3. Total installed wind power capacity in the world in megawatts until 2007
and the prediction for 2008-2010. Data provided by the World Wind Energy
Association.
5
Globally, in 2007, the total installed wind power capacity is roughly 94000
MW and, as can be seen in Figure 1.3, the prediction for the next years is to almost
double that capacity.
In the Netherlands, the installed wind power capacity currently is

approximately 1500 MW. The goals of the Dutch government are to increase this
capacity to 3000 MW in the next few years. In 2020, the goal is to have 20% of the
total output power be generated by means of sustainable energy, which for a
significant part will consist of wind energy. This will mean that, by then, a total
wind power capacity of 6000 MW will have to be installed, see [8].
1.3.1 Conventional methods
The most common and well known method of converting wind energy to
electrical energy is through the use of wind turbines of the horizontal axis design,
see Figure 1.4.
Figure 1.4. Wind turbines in the middle of the Dutch farming landscape.
Their current efficiency is around 50% at their rated speeds, which means
they convert 50% of the available power associated with the wind. Currently,
smaller turbines have power capacities around 250 kW, while the larger turbines
are capable of delivering several megawatts to the grid.
Another rotation based method is the Darrieus wind turbine, shown in

Figure 1.5, which is based on a vertical axis instead of the horizontal axis type that
is used in the conventional wind turbine. This design has the advantage that it
utilises wind from all directions.
6
Figure 1.5. An example of a Darrieus wind turbine.
1.3.2 Drawbacks of wind turbines
So, while wind turbines still are the main devices to convert wind energy to
electrical energy, there are a number of drawbacks that limit the widespread use of
wind energy. The main drawback is the high cost of maintenance. This need for
maintenance arises primarily from the conversion of wind energy to electrical
energy via mechanical energy, i.e. the rotational movement that drives the wind
turbine. Especially gear box driven wind turbines are prone to wear and tear and
need to be maintained on at least a yearly basis. Added to the cost of maintenance
are the costs for, amongst other things, construction, land lease and permits which
makes government subsidies a requisite to enabling wind energy projects. Thus, at
this point, the choice for the wind energy is a political one [9].
Other drawbacks include the fact that conventional wind turbines are
bound to circular surface areas, because of the rotational movement. This rotational
movement is also the cause of noise and intermittent shadow nuisance. Another
often heard complaint is that these wind turbines are responsible for what is called
“visual pollution”, especially when large wind turbine farms in rural settings are
involved. A solution for this problem is to build wind turbine farms at sea. This, of
course, introduces problems such as increased construction and maintenance costs.
7
1.3.3 Alternative wind energy methods
There have been other developments in the field of alternative wind energy
conversion. One example is the so-called “ladder mill” [10] which consists of a
series of kites moving in a rotating manner while driving a generator.
Thus far, all of these developments are still in an experimental stage and
have not been yet proven commercially successful. Also, they all have the common
element that there are mechanically moving parts present in the design, which will
lead to wear and tear, similar to wind turbines.
1.3.4 Use of wind energy
As many countries have stated in their long-term energy goals that they
wish to increase their wind energy output, a question that often arises when any
form of alternative energy production is mentioned is whether it can replace current
energy production processes. With respect to wind energy, it can be said that, due
to its intermittent nature and the limited means of energy storage in general, in the
foreseeable future wind energy will not be able to fully replace fossil and nuclear
based energy production.
Also, computational research using meteorological models as described in

[11] suggest that the local and global climate could be altered by the use of large-
scale use of wind energy, for example 1/10th of the global electricity demand, by
extracting kinetic energy and changing turbulent transport in the atmospheric
boundary layer.
1.4 A new concept, a system with very little

mechanical movement
As stated before, in all of the methods, that are used to convert wind energy
into electrical energy, some form of mechanical movement occurs, which is the
primary reason for maintenance and usually the primary cause of failure.
Therefore, a concept in which there is very little mechanical movement would be
ideal with respect to system complexity and maintenance costs.
The EWICON method (Electrostatic WInd energy CONverter) is a method

which is based on the principle that the wind transports electrically charged
particles or charge carriers in an electric field. Without going into great detail on
how we will go about creating these charge carriers (this will be discussed in
chapter 3), for now, we will discuss the principles of the EWICON method and its
8
possible implementations into an actual system. In principle, any object, that can
hold or store a charge, could be used as charge carrier. At the end of this chapter,
we will discuss in more detail what this, in practice, will come down to.
It is important to mention that the concept of converting wind energy into

electrical energy by having the wind move charge carriers in an electric field is not
new. There have been several patents that propose a similar idea, see for example
[12] and [13]; however, most of these patents do not have an energy efficient
solution for creating charge carriers.
1.4.1 Principle: work on charge carriers
When a force acts on a body that undergoes a displacement, that force does
work on the body. In the case of the EWICON system, the body is a charged
particle with a charge q and the force is the electric force on the charged particle
due to an electric field E, given by
G G
F = q⋅E (1-1)
which is illustrated in Figure 1.6.
Figure 1.6. A positively charged particle is pushed towards the positive electrode by
the wind against the direction of the electric field, thereby increasing the potential
energy of the charged particle.
By allowing the wind to force the charged particles against the direction of
this electric force, the potential energy of these charged particles will increase,
similar to pushing a rock up a mountain against gravity.
9
dW = − dU (1-2)
These charged particles with increased electrical energy can then be collected using
one of the methods explained in the next section. In this way, wind energy is
directly converted into electrostatic energy and the intermediate step involving the
rotational movement, which takes place in conventional wind turbines, is taken out
of the process.
1.4.2 Implementation of the EWICON method
Currently, there are two methods of collecting the charged particles, both
of which will be explained together with their respective advantages and
disadvantages:
1.4.2.1 Type A: the patent of Alvin Marks et al.:
As we can see in Figure 1.7, in this implementation of the EWICON

system, the charged particles are created by a charging system, which usually
consists of a number of nozzles and electrodes, which is grounded. A stream of
charged particles, which can be considered as an electric current, is then
transported by the wind to a separate insulated collector, which is initially neutral.
When the charged particles touch the collector, they will deliver their charge to the
collector. This causes the potential of the collector to rise.
This potential will have the same polarity as the charged particles cloud,
thereby creating an electric field. Due to this field, an electric force will push the
charged particles away from the collector. Initially, the wind force will be larger
than the electric force and therefore the charged particles will still arrive at the
collector. As long as this process occurs, however, the electric field generated by
the collector will continue to increase, causing the charged particles cloud to either
move back against the wind or around the collector. When the charged particles
come in contact with the charging system or earth, the charge is lost and the net
current decreases.
Therefore, the wind has to overcome this repelling electric force and
depending on the speed of the wind, the size of the collector and the load, the
collector potential attains a maximum that further depends on possible leakage
currents to earth via the insulator surface. If all produced charged particles are
captured by the collector, then the maximum power of the EWICON has been
attained.
10
Figure 1.7. An example of a Type A implementation of the EWICON system with an

insulated collector. The charged particles are created at the charging system
consisting of nozzles and electrodes. The wind transports these particles to the
collector. In previous experiments, the distances between nozzle and collector were in
the order of one to two metres. A load can be attached to the collector.
Currently, research on this topic is also carried in Japan by Sato et al., see
[14]. This work is only limited to implementations with a separate collector.
1.4.2.2 Type B: the collector-less EWICON system:
In this implementation, depicted in Figure 1.8, the charging system itself is

insulated from earth. There is no separate collector present. Since the charging
system starts in an electrically neutral situation, dispersing charged particles will
cause the potential of the charging system to rise. To be more precise, the potential
will rise in case of negatively charged particles, it will decrease in case of
positively charged particles. However, this time, the polarity of this potential is
opposite to the polarity of the charged particles. This means that in the absence of
wind, the charged particles will be forced back to the charging system resulting in a
charge loss or net current decrease.
11
Again, depending on the speed of the wind, the quality of the insulation of
the charging system and the load attached to the system, the system itself will reach
a maximum potential. Basically, the earth acts as the collector for the charged
particles and if all of the charged particles are transported to earth, then the
maximum power that can be delivered is attained.
Figure 1.8. An example of a Type B implementation of the EWICON system. Here,

earth acts as the collector. The charging system itself is insulated from earth and as a
consequence, the dispersal of charged particles will result in the rise of the potential of
the system. Again, a load can be attached to the charging system.
Also, in this case, an electrical load can be connected to the charging

system.
1.4.2.3 Comparison of both implementations
Although the two implementations are equivalent from an electrical point

of view, practical implications show that there is a preference for the Type B
implementation. While the Type A implementation is conceptually simpler, it does
require an external collector, which means extra building efforts. The drawback of
a Type B implementation is that the charging system needs to be isolated from
earth, which significantly increases the complexity of the EWICON. However,
because there is no need for an external collector, the overall EWICON design
12
becomes simpler, effectively making the EWICON a one-unit system. This, for
example, would make it relatively simple for the EWICON system to be aligned to
the direction of the wind, whereas the working direction of a Type A
implementation would be more or less permanent after building and placing the
collector and the charging system. A platform could be designed on which both the
collector and the charging system could be placed. In this way, the whole system
could be aligned with changing wind directions.
Also, in the case of an external collector, due to the inherent varying

character of the wind, a relatively large collector needs to be built to ensure that
most of the charged particles are collected and even then, there is no guarantee that
all of the particles will be collected. The percentage of collected particles depends
on the size and shape of the collector. However, when considering Type B, the
only requirement is that the charged particles do not return to the charging system.
In this way, the rest of the surrounding environment acts as a collector.
A Type A implementation, by its nature, would lend itself for a (partial)

particle recycling system, thereby reducing the amount of liquid used for energy
production. This is an important issue to keep in mind when considering the Type
B implementation, where the charged particles are dispersed into the surroundings.
Therefore, from an environmental point of view, the charged particles cannot be
harmful or toxic. One should note, however, that this problem would also be
present when using the Type A implementation, since, as stated in the previous
section, it would require a very large collector to catch all of the dispersed
particles. Even then, there will inevitably be particles that would not reach the
collector because of wind speed and direction and because of the fact that the
collector would repel particles.
Table 1.1. Comparison of the two implementations of the EWICON system.

Type A Type B
Charging system Grounded Insulated
External collector Yes No
Extra building effort Yes No

Only with both charging
Alignment with changing
system and collector on a Yes
wind direction
single rotational platform
13
1.4.3 Expected advantages and disadvantages
The EWICON principle is expected to have several advantages over the

standard conventional wind turbine systems, also listed in Table 1.2.
First of all and most importantly, apart from the floating charged particles,
there are no moving/rotating parts present in the EWICON system. This means that
wear and tear commonly found in the gearbox systems of wind turbines will not be
present in the EWICON system. This will have a positive effect on the
maintenance and investment costs.
Secondly, due to the lack of moving/rotating parts, there is less noise

originating from the EWICON system. The noise would be comparable to that of
wind flowing around and through tall buildings or structures.
Thirdly, since there are no rotating blades present, no intermittent shadows

will be present. This is especially important for the placement of wind energy
systems in the urban or rural settings.
Finally, since there is no rotational movement in the EWICON system,

increasing the wind surface area does not necessarily have to go hand in hand with
an increased circular wind area, as illustrated in the Figure 1.9.
Figure 1.9. Using conventional wind turbines (left), increasing the wind surface area
means the diameter of the circular area needs to be increased. Using an EWICON on
for example a dam (right), the wind surface area is increased by stretching the width
of the EWICON.
14
Due to the nature of the EWICON system, it is possible to extend the wind
surface in one of the two dimensions, making a modular approach possible. This
way, one could think of building long strips of EWICON modules along rooftops
or dams, as illustrated in Figure 1.9. These long strips would, however, be static
constructions and thus, less adaptable to the direction of the wind.
There are some issues associated with the conventional use of wind energy
that cannot be solved by the EWICON system. We will see in the next chapter that
the wind surface area is a restrictive parameter for the maximum of extracted
power. This means that the EWICON system will have dimensions comparable to
those of conventional wind turbines which will not solve the “visual pollution”
problem due to size. However, because the construction and maintenance costs are
expected to be lower than those of wind turbines, the placement of EWICON
systems at sea might eventually be commercially more favourable than the
placement of wind turbines farms.
In the introduction of this section, we briefly mentioned that, in theory, the

charged particles could be any type of object that is able of storing electrical
charge. However, in later chapters, we will see that using liquids particles or
droplets as charge carriers currently is the most practical way of having the
disposal of streams of charged particles. Considering the fact, then, that the
EWICON system is based on liquid particles and that they are charged by
electrodes, long term operation in humid and possibly saline environments, like at
sea, might result in other problems than the usual wear and tear, like corrosion of
the electrodes, electrolysis and clogging of liquid supply systems.
Finally, as we will see in chapter 3, creating the charged particles will

require certain electrodes to operate on high voltage, see [15]. Therefore, attention
must be paid to safety regulations and insulation issues. These problems will have
to be addressed and taken into account when making the cost/benefit analysis.
Table 1.2. Comparison of an EWICON system with most conventional wind turbines.
wind turbines EWICON
Rotational mechanical
Yes No
movement
Comparable to noise of air
Noise Yes
flow along static buildings
Intermittent shadows Yes No
Linearly, in both directions

Modular expandable Somewhat
to a degree
15
1.4.4 Goals EWICON research project
The goals of this PhD research project were to investigate the theoretical
and practical issues that have to be considered when designing and building an
actual system based on the EWICON principle with both Type A and B
implementations in mind. The research, needed to ultimately build an EWICON
system, has been approached in the following manner:
¾ The dependence of the efficiency of an EWICON system on

parameters like wind speed, type of liquid, droplet size and charge
was determined, thus establishing a model that is able to provide
an estimation of the output power.
¾ Suitable methods for creating the charged droplets were

investigated and tested whether they are compatible with the
requirements of an EWICON system.
¾ The electrical behaviour of the EWICON system due to the high

voltage components was analysed in order to understand the
interaction between the charged droplets and the power output of
the system.
1.4.5 Contents and approach
Clearly, the stage of development of the EWICON is very early, in a sense

that at the beginning of the current research, the proof of the principle had to be
confirmed in practice.
First of all, the theoretical and practical feasibility of an EWICON system

will be discussed. In chapter 2, an elementary model of the system, incorporating a
number of charged droplets in an electric field and wind, will be developed which
can be used to calculate the expected power output. With this model, we can then
determine whether the EWICON system is able to achieve a net power output.
Concurrently, a method must be devised to calculate the efficiency of an EWICON
system based on the available wind energy, the energy needed to operate the
EWICON system and the liquid consumption. This way, various implementations
of the EWICON system can be compared.
Secondly, in chapters 3 and 4, based on results of simulations, both of the

developed model and field analysis/droplet trajectories, a charging system has been
designed. This will include the size and shapes of various electrodes, the geometric
configurations, voltage potentials at which these electrodes need to be set. Also,
16
using the simulations, it will be obtained at what wind speeds the designed
EWICON system is expected to function. Since charged droplets play an important
role in this system, a method is needed to create droplets and to apply electrical
charge to these droplets. Several methods will be discussed and analysed in terms
of energy efficiency and liquid consumption.
Finally, in chapters 5 and 6, a laboratory version of an EWICON system

will be built and tested and suggestions will be given for an industrial prototype
and for a scaled up version. Also included in chapter 5, some consideration is given
to the social, economic and medical issues surrounding the EWICON system.
1.5 Acknowledgment
The author wishes to acknowledge Dr. Piet Sonneveld of the Wageningen

University and Research Centre for reintroducing the EWICON concept and for the
close cooperation with the research project.
17
Chapter 2 – Theoretical basis and modelling
2
Theoretical basis and modelling
“In theory, theory and practice are the same. In practice, they’re not.”
– Lawrence Peter “Yogi” Berra
In this chapter, to help understand the basic principles of this concept, we

will construct a model which we can use to make predictions of the operation and
the efficiency of the EWICON system. This will involve assessing the power
associated with the wind and how much of it can be used for the conversion. Based
on the movement of the charged carriers under influence of the electric field and
the wind, we can determine whether the carriers will be removed from the system
and if so, how much electrical energy potentially can be gained. Using this model,
we can determine which parameters are the most important for the efficiency of the
EWICON system.
This model will be used as the basis for computational simulations.

Quantitative results from this model can then be compared to experimental results
and used to further improve and optimise EWICON system parameters.
Practical issues like the energy costs of all the equipment or devices needed
for e.g. liquid supply or high voltage sources for the electrodes will not be taken
into account during these theoretical considerations. All these issues will be
addressed in chapter 4 and 5 where we will discuss the practical design of an
EWICON system and the experiments conducted with the EWICON system.
At the end of this chapter, we will discuss in what way the efficiencies of
various EWICON implementations will be determined. We will compare these
implementations by defining a benchmark called the EWICON Performance Index
or EPI. In the EPI, the energetic costs of the equipment and devices will be taken
into account.
19
2.1 Energy and power contained in the wind
The kinetic power, Pw, contained by a flow of air is related to the 3rd power
of its speed, vw,
Pw = 12 ⋅ A ⋅ ρ a ⋅ v w3 (2-1)
in which A is the active surface area perpendicular to the direction of the

wind flow and ρa is the air density. Not all this power can be used for the
conversion of wind power to other forms of power because this would imply that
the air flow would come to a complete stop just behind the converter.
This has been represented in Figure 2.1, where the original wind speed, vw
as stated in equation (2-1) is indicated as v1. The speed of the wind after it has
passed through the converter has been indicated as v2.
Figure 2.1. Illustration of the Betz limit. The wind speed behind the wind converter,
v2, should ideally be ⅓ of the speed before it enters the converter, v1, while the speed
just in front of the converter is ⅔ of the initial wind speed.
By determining to what extent the wind would have to be slowed down by

the converter and by calculating what the final wind speed would be, Betz [16] and
Lanchester [17] both derived an expression for the maximum recoverable power,
Pmax, from the wind as a function of the prevailing wind speed vw:
20
Pmax = 8
27 ⋅ A ⋅ ρ a ⋅ v w3 (2-2)
In deriving expression (2-2), the wind speed, just in front of the converter,
is ⅔ of the original wind speed vw and far behind the converter the wind speed is ⅓
of the original wind speed vw. A derivation of this expression can be found in
appendix A. This distribution of wind speeds is the best compromise between using
the power of the wind and maintaining a flow of air. The Betz limit basically states
that there is an upper limit to the amount of power that can be extracted from the
wind. More important, this limit is irrespective of the method of converting, which
means that the EWICON principle is also bound by this limit.
Looking at the ratio between (2-1) and (2-2), we find that this limit is
Pmax 16
= ≈ 0.593 (2-3)
Pw 27
which basically states that this upper limit is almost sixty percent. The net
power produced by the EWICON or any wind converting system for that matter
will in part be assessed with respect to this limit.
2.2 EWICON model
In this section, we will look at the various aspects involved in the

modelling of the EWICON concept. Such a model will consist of a number of
charge carriers moving in an external electric field while different forces,
mechanical and electrical, will act on these charge carriers.
Firstly, the charge carriers will be defined in terms of type, charge and size.
Secondly, all the relevant forces will be specified, which, after calculations, will
yield the trajectories of the charge carriers, i.e. positions and velocities. Finally, we
will look at the constraints on the droplet parameters, the electric field and the wind
speed, which must be observed in this model in order for the wind to transport the
droplets away.
2.2.1 Rayleigh limit and droplet parameters
First of all, charge carriers have to be defined, which according to IEC

norm no. 114-14-44 are “particles having one or more elementary electric
charges”. In the previous chapter, we stated that, in theory, any object that can be
wind-driven and hold an electrical charge is a useable charge carrier. However,
based on the charge creation methods which will be discussed in the next chapter,
21
it turns out that using a liquid to create small droplets is a convenient way of
creating charge carriers.
In this case, if a liquid is to be used as the source of charged droplets, there

is a limit to the amount of charge that can be present on such a liquid droplet. This
is because of the mechanical instability that will occur when the surface tension of
the liquid and the electrostatic stress due to the repulsion of charges at the droplet
surface are equal. If the charge on the droplet becomes too large, then the droplet
will break up into smaller droplets. This occurs at the Rayleigh limit, which
depends, among other parameters, on the droplet size. Therefore, the droplet size,
and the ability to control it, will be important factors to consider when discussing
methods for charged droplet creation in the next chapter.
This Rayleigh limit or the maximum charge, qmax, that can be present on a
droplet just before it breaks up, is given by
q max = 2π 2γ ⋅ ε 0 ⋅ d 3 (2-4)
in which γ is the surface tension of the liquid of the droplet, ε0 is the vacuum
permittivity and d is the droplet diameter.
In Table 2.1, we can find values of qmax corresponding to a number of

droplet diameters, for two types of liquids, water and ethanol, with γwater = 72·10-3
N/m and γethanol = 22·10-3 N/m. As we can see, for each droplet diameter, the qmax is
lower for ethanol than for water due to the lower surface tension of ethanol.
Table 2.1. Example values of the maximum charges on water and ethanol droplets as
specified by the Rayleigh limit as a function of the droplet diameter, with γwater =72·10-3
N/m and γethanol =22·10-3 N/m.
Droplet Maximum Charge to Maximum charge Charge to
diameter charge on droplet mass ratio for on droplet for mass ratio for
(μm) for water (C) water (C/kg) ethanol (C) ethanol
(C/kg)
1000 2.26·10-10 1.30·10-2 1.26·10-10 9.10·10-3
100 7.15·10-12 4.10·10-1 4.00·10-12 2.86·10-1
10 2.26·10-13 1.30·101 1.26·10-13 9.06
-15 2 -15
1 7.15·10 4.10·10 4.00·10 2.86·102
0.1 2.26·10-16 1.30·104 1.26·10-16 9.06·103
22
Also, for each droplet diameter, the maximum charge to mass ratio, CMR,
has been calculated and included in Table 2.1, as given by
q max
CMR = (2-5)
mdroplet
This ratio illustrates the fact that, while the charge on a single droplet in
absolute terms becomes smaller with decreasing droplet size as shown in Figure
2.2, the relative charge on a droplet increases significantly as shown in Figure 2.3.
-11
10
-12
10
maximum charge (C)
-13
10
-14
10
-15
10
-6 -5 -4
10 10 10
droplet diameter (m)
Figure 2.2. The maximum possible charge on a droplet as specified by the Rayleigh
limit depicted as a function of the droplet diameter, ranging from 1 to 100 μm. Graphs
are shown for water (continuous line) and ethanol (dashed line).
This means that with equal amounts or equal flow rates of a liquid, more
electrical charge can be produced. Therefore, if all other factors are equal, the
droplet diameter should be as small as possible. However, before we can say that
with certainty, we will investigate in the next few sections, where we will discuss,
amongst other things, constraints on droplet parameters, whether there are certain
restrictions to the droplet diameter.
The maximum charge on a droplet as specified by the Rayleigh limit

should also be carefully watched with respect to the evaporation of the liquid
23
droplet. Due to this evaporation, the droplet diameter will decrease over time. One
of the main parameters governing the rate of evaporation is the difference between
the velocities of the droplet and the surrounding air. As the droplet is moved along
by the wind, the difference of droplet velocity with that of the wind decreases and,
thus, does the rate of evaporation. This difference will be the greatest at the
moment the charged droplet is created and released from the charging system into
the moving air. Therefore, the rate of evaporation will be at its highest just as the
charged droplet is starting its flight.
2
10
1
10
charge to mass ratio (C/kg)
0
10
-1
10
-2
10
-3
10
-6 -5 -4
10 10 10
droplet diameter (m)
Figure 2.3. The charge-to-mass ratio, CMR, depicted as a function of the droplet
diameter, ranging from 1 to 100 μm. Graphs are shown for water (continuous line)
and ethanol (dashed line).
If the rate of evaporation is sufficiently high, the charge that is present on

the shrinking droplet will reach the Rayleigh limit and, at this point, the charged
droplet will break up into smaller droplets. If the rate of evaporation continues to
be high, then the smaller charged droplets will continue to break up until only ions
are left.
This break-up process is illustrated in Figure 2.4. The wind drag force,
which among other things is dependent on the droplet diameter, decreases, while
the opposing electric force remains constant. In other words, the electric mobility,
which will be discussed in 2.2.3, is very high. Therefore, if the ions are still
24
sufficiently close to the charging system, the wind will not be able to transport the
electrical charge that was initially stored in the original droplets.
Figure 2.4. The break-up of charged droplets due to evaporation effects in

combination with the maximum charge as specified by the Rayleigh limit. a) The
droplet diameter decreases due to evaporation, b) the droplet breaks up into more
droplets, c) due to continued evaporation, the newly formed droplet keep shrinking
and breaking up until d) only ions are left.
Preferably, the charged droplets should survive as long as possible until

they have reached the collector or the earth. Therefore, it is important to calculate
the evaporation times, i.e. the time between state A and state D in Figure 2.4, of the
charged droplets while they are being transported by the wind. In the next section,
an analytical model will be derived, which can be used to calculate the flight times
of the charged droplets. Comparing the evaporation times with flight times, we can
determine whether the charged droplets can contribute to the conversion process.
Thus, in order to calculate the evaporation times of water droplets, the rate
of evaporation has to be determined. This can be achieved in several ways, e.g. by
using an energy balance method or by using mass transfer/diffusion methods. In
this research, the latter method has been used which resulted in Table 2.2.
See appendix B for the set of equations and empirical relations that have
been used to calculate the values in Table 2.2. The evaporation times of water
25
droplets with diameters varying from 1 to 100 μm have been calculated at a

temperature of 20 °C.
Table 2.2. The evaporation times of a water droplet as a function of the droplet
diameter. Also, the relaxation time, τ, of the droplet has been calculated at 20 °C and a
relative humidity of zero percent.
droplet diameter, dd evaporation time relaxation time, τ
(μm) (s) (ms)
1 0.001 0.003
10 0.076 0.304
20 0.31 1.21
30 0.69 2.73
40 1.22 4.86
50 1.91 7.59
60 2.75 10.93
70 3.74 14.88
80 4.89 19.43
90 6.19 24.59
100 7.64 30.36
In this table, we can see, for instance, that for droplets in the 20-50 μm
range, the evaporation times vary from 310 milliseconds to almost two seconds if
there is no wind present. If the air surrounding the droplet is moving, the
evaporation time will be shorter depending on the difference between the speed of
the droplet and the speed of the surrounding air. This means that water droplets
need to be sufficiently far away or possibly collected by the earth before they are
fully evaporated. These calculations have been conducted assuming that the
relative humidity is zero percent.
In order to fully assess whether the evaporation will complicate the

movement of the charged water droplets, we also have to consider the relaxation
time, τ, of the droplets. In general, the relaxation time is the time required for a
droplet to adjust to an equilibrium condition from a non-equilibrium position. In
this context, it is the time it takes for the charged water droplet to attain the same
speed as the surrounding medium i.e. the wind. As soon as that is the case, we only
have to take into account the evaporation times as listed in the 2nd column of Table
2.2.
In appendix B, the reader can find the procedure for finding the relaxation
times, but we will suffice by stating that these times for water droplets from 1 to
100 μm have relaxation times varying from 3·10-6 to 30·10-3 seconds, as can be
seen in Table 2.2. In practice, this means that compared to the evaporation times,
26
the relaxation times are very small. This, in turn, implies that for practical matters
we can take the 2nd column of Table 2.2 as the characteristic evaporation times
which we can compare with the flight times. These flight times will be calculated
in the next section and compared to the evaporation times.
2.2.2 Analytical model for simple configuration
In order to model the movement or trajectories of a number of the charged

droplets as described at the beginning of this section, these droplets will be
positioned in a simple electrode configuration, which means that the electric field
will be modelled as a uniform field decreasing in the wind direction with 1/r2. The
droplets will be assigned a polarity such, that the electric force they will
experience, will be directed in the opposite direction of the wind, as shown in
Figure 2.5.
First of all, equations must be given for the model, as complete as possible
within reason.
Figure 2.5. The schematic representation of the basic analytical model of charged
droplets in a uniform electric field in which there is also wind present. Also, these
charged droplets repel each other because of the like polarity.
27
There are several forces acting on the droplets, starting with gravity which
obviously acts on all droplets,
G G
Fi , g = mi ⋅ g (2-6)
in which mi is the mass of the ith droplet and g is the gravity acceleration. In
this model, for the time being, we will keep the mass of the droplet, mi constant,
thus disregarding evaporation effects. After droplet flight times have been
calculated, we can compare them to the life times of the droplets and determine
whether evaporation is a critical factor.
Because these droplets fall in air, there is also an upwards force acting on
the droplets, the buoyancy,
G G
Fi , B = − ρ a ⋅ Vd ⋅ g (2-7)
in which Vd is the volume of the droplet and ρa is the air density. Next,
there is the drag force due to the wind which initially causes the droplets to move,
G G G
π G G 2 v −v
Fi , w = ⋅ CD ⋅ ρ a ⋅ d 2 ⋅ vw − vd ⋅ Gw Gd (2-8)
8 vw − vd
in which CD is the drag coefficient, vw is the wind speed and vd is the speed
of the droplet. This drag force will be zero when the droplet moves with the same
speed as the wind. In specific cases, we can take out CD and simplify (2-8) by
looking at the Reynolds number.
The Reynolds number is an important dimensionless number in fluid

mechanics and it is used to determine whether a flow process is turbulent or
laminar by looking at the ratio of inertial and viscous forces. If the Reynolds
number is lower than 1, which means that the viscous forces are dominant, then the
flow process is said to be laminar,
G G
ρ a ⋅ d ⋅ vw − vd
Re = <1 (2-9)
ηa
in which ηa is the dynamic/absolute viscosity of air. Laminar flow is often

associated with smooth flow patterns as opposed to turbulent flow like e.g. the air
flow around an airplane wing. When this happens, it is said that the flow process
28
takes place in the Stokes region and it can be shown that, in that case, Stokes’ law
can be used for the drag force,
G G G
3π ⋅ η a ⋅ d ⋅ (v w − v d )
Fi , w = (2-10)
Cc
in which Cc is the slip correction factor by Cunningham, which becomes

relevant for particles smaller than 1 μm. For particles with d >> 1 μm, we can
assume Cc = 1, see [18].
The electric forces acting on the droplets can be divided into two parts. The
first part is the electric force due to the external electric field created by one or
more electrodes present in the EWICON system, which will be represented by Eext,
G G
Fi , E = qi ⋅ E ext (2-11)
in which qi is the charge present on the ith droplet. This charge will be
expressed as a percentage of the Rayleigh limit and its value will depend on the
spraying method used to create the charged droplets. The electric field, Eext, is
assumed to be decreasing in the wind direction with 1/r2 and pointed in the
opposite direction of the wind. In this modelling, part of Eext represents the field
created by the charging system of the EWICON and, thus, the field that is
responsible for the creation of the droplets.
The second part of the electric force due to the electric fields created by the
other charged droplets is,
G 1 qi q j
Fi , j = ⋅ ⋅ rˆij (2-12)
4πε 0 ri 2,j
in which ri,j is the distance between droplet i and droplet j. It should be

noted that all the created charged droplets have charge with the same polarity and,
thus, they will repel each other.
In reality, the sum of the electric fields generated by all the charged
droplets affects the electric field at the charging system, represented by Eext, and
that means that the charging and spraying process is changed. This, in turn, would
mean that the created charge droplets could have different sizes, initial velocities
and charges. In this model, we will not take this into account.
29
Summing all contributing forces, we find for the law of motion
G G G G G j G
G
Fi = Fi , g + Fi , B + Fi , w + Fi , E + ∑ Fi , j = mi ⋅ ai (2-13)
j ≠i
in which Fi is the total force on the ith droplet and ai is the acceleration of
the charged droplet. Equation (2-13) can be solved, thereby giving us the velocity
and position of every droplet at each point in time. In this equation, if we look at
the orders of magnitude, then the total force is mainly dominated by the drag force
and the electric force.
Using (2-13), the work done on the ith droplet, Wi, by the wind can be
found by using
G G G
( )
Wi = ∫ Fi − Fi , w ⋅ d A (2-14)
in which dℓ is the displacement which follows the path of the droplet. From
this, the potential energy difference for the droplet can be calculated and, when
taking the sum of all the droplets, the total energy gained from the wind.
A number of assumptions have been made using this representation of the

movement of charged droplets by wind in an electric field. Firstly, as has already
been mentioned, we assume that there is no evaporation of the droplets. If these
charged droplets do evaporate, then ions will be formed and there will be an ionic
space charge field. If this occurs close to the charging system of the EWICON,
then the influence of this field should be taken into account. Therefore, the
evaporation times calculated in the previous section should be compared to the
flight time of the charged droplets in order to verify whether the omission of
evaporation effects is justified.
However, estimations show that in wet conditions the evaporation times
are in the order of ten seconds. With a wind speed of e.g. 10 m/s, this implies that
the ionic space charge field will be formed at distances of a hundred metres and its
effect can be neglected.
Secondly, we assume that the size distribution of the droplets is

monodisperse, i.e. all the droplets have the same size, and that the droplets each
have the same charge. This, again, depends on the used spraying method and the
way it has been implemented in the EWICON system. Thirdly, we assume that the
droplets remain spherical at all times and do not deform when they are moved by
the wind.
30
Finally, with respect to the wind speed, we assume that the flow of air is
constant, laminar and unidirectional. In realistic scenarios, this is not always the
case. However, during the experiments with similar systems, which will be
discussed in chapter 5, the flow of air could be controlled such that it satisfied these
assumptions.
2.2.3 Constraints on size and charge, electrical mobility
With the knowledge of the maximum charge, qmax, as described in section

2.2.1, we have to choose a certain value for the charge, q, on the droplet, which we
will express in percentages of the Rayleigh limit. In practice, this will depend on
the chosen charge creation method and the mode in which this method is operated.
Some of these methods have the possibility to independently control the charge on
the produced droplets, while other methods have size and charge dependency, see
chapter 3 on methods for charge droplets creation for more information.
Regardless of which of the two implementations discussed in chapter 1 is

used, for the EWICON system to operate successfully, the charged droplets need to
be transported away by the wind completely. This flow of charge constitutes an
electric current. The more charge can be transported per time unit, the higher the
electric current will be and, in turn, the higher the output power will be, if we
assume that the working voltage of the EWICON system remains constant during
operation.
Therefore, it may seem favourable to assign as much charge to a droplet as

possible and to make the droplets as small as possible, resulting in a high charge-
to-mass, CMR, ratio, because in that case more droplets per time unit can be
dispersed using the same liquid flow rate. However, these small highly charged
droplets may have trouble to overcome the electric field which attracts them back
to EWICON system and thus the question arises of what the optimal size of and
charge on a droplet should be. Related to the answer to this question, it should also
give an indication of the maximum of the electric field and/or the minimum wind
speed needed to transport the droplets away.
If a charged droplet is released in a uniform electric field in vacuum, then

the charged droplet will experience a constant electric force and, thus, will move
with increasing speed, as indicated in Figure 2.6.
31
Figure 2.6. A charged droplet is released in a uniform electric field in vacuum. The
droplet experiences a constant electric force to the left and will be constantly
accelerated.
If a charged droplet is released in a uniform electric field in stationary air,

then the droplet will experience not only a constant electric force, but also a drag
force in the opposite direction, as indicated in Figure 2.7. This drag force is
dependent on the speed difference of the droplet with the surrounding air.
Therefore, the drag force will increase as long as the droplet is accelerated by the
electric field until both forces are equal in magnitude. At that point, the droplet will
move with a constant speed to the left.
Figure 2.7. Here, a charged droplet is released in a uniform electric in stationary air.
Now, the droplet will also experience a drag force, Fw, in the opposite direction,
dependent on the difference between the speed of the droplet and the surrounding air
and at the certain moment the two forces will be equal.
32
This situation is similar to that occurring at the nozzle of an EWICON

system, where droplets are released in moving air flowing from the left to the right,
as indicated in Figure 2.8.
Figure 2.8. The situation in an EWICON system. A charged droplet is released in

wind moving from the left to the right. In order for the wind to be able to move the
droplet to the right, the wind speed needs to be higher than the constant speed that
was reached by the droplet in Figure 2.7.
If the wind is to move the charged droplet to the right, then the wind speed
must be higher than the constant speed or terminal electrostatic velocity, vT, that
was reached in stationary air in Figure 2.7. In order to calculate this terminal
electrostatic velocity, we look at the mechanical mobility, μm, and the electrical
mobility, μe, of the charged droplets which are defined as:
vT
μm ≡ (2-15)
Fw
and
vT
μe ≡ = q ⋅ μm (2-16 )
E ext
in which vT is the terminal electrostatic velocity of the droplet or, in this

case, the difference between the wind speed and droplet speed, |νw - νd|. Fw is the
wind drag force and Eext is the external electric field, in this case, generated by the
electrodes present in the EWICON system. The latter part of equation (2-16) is
derived when we equate the electric force to the drag force, which can be expressed
in the form of equation (2-10) if the droplets move in the Stokes region.
33
If the Reynolds number is larger than 1, then the expression for the drag
force becomes a bit more complicated. Again, we equate the electric force to the
drag force, but this time, we use equation (2-8) for the drag force.
π
q ⋅ E ext = ⋅ C D ⋅ ρ a ⋅ d 2 ⋅ vT2 (2-17)
8
in which |νw - νd|2 has been replaced with νT2. Equation (2-18) can, if we
solve for CD and multiply by Re2, be rewritten to
8 ⋅ q ⋅ E ext ⋅ ρ a
C D ⋅ Re 2 = (2-18)
π ⋅η 2
On page 57 of [18], we can find that for the terminal electrostatic velocity
the following expression can be given
⎛ η ⎞ 2
vT = ⎜⎜ a ⎟⎟ ⋅ e ( −3.070+ 0.9935⋅ J −0.078⋅ J ) (2-19)
⎝ ρa ⋅ d ⎠
in which
(
J = ln CD ⋅ Re 2 ) (2-20)
For a range of droplet diameters and charges, one can calculate the
terminal electrostatic velocities, vT, of the droplets as a function of the electric
field. This will give an indication whether the droplets have a too small or too big
diameter or too much charge such that the wind is not able to transport them away.
Also, this will give an indication on how to shape and grade the electric field to
make sure it is not too high for the charged droplets to escape from. This is a point
that will be addressed in detail in chapter 4.
As an example, in Table 2.3, we have calculated the maximum charge for

four water droplets, each with a different diameter. In this table, for now, we have
set the charge on the droplets to 70% of the Rayleigh limit maximum charge. In the
next chapter on charged particle creation methods, we will see that there are
methods available that can achieve these kinds of charging efficiencies.
34
Table 2.3. Various water droplet diameters with their respective maximum charge. In
this example, the charge on the droplets, q0.7, is set to 70% of the maximum charge
which corresponds to charging efficiencies of one of the charged particle creation
methods. These droplets are used to calculate the terminal electrostatic velocity in an
electric field and air.
d (μm) d (m) qmax (C) q0.7 (C)
-7 -16
0.1 10 2.25·10 1.58·10-16
1 10-6 7.12·10-15 4.98·10-15
10 10-5 2.25·10-13 1.58·10-13
100 10-4 7.12·10-12 4.98·10-12
Continuing the example, while assuming that the flow process takes place
in the Stokes region, we calculate the terminal electrostatic velocities for one of the
droplet diameters mentioned in Table 2.3. In this example, we take 1 μm as the
droplet diameter and we perform the calculations for different electric field
strengths, of which the results can be found in Table 2.4.
Table 2.4. Terminal electrostatic velocities, vT, for increasing electric field strengths.
droplet diameter E (V/m)
2
1 μm 1 10 104 106
Stokes vT (m/s) 3.40·10-5 3.40·10-3 3.40·10-1 3.40 101
2.24
Re 2.24·10-6 2.24·10-4 0.02
too high ↓
Newton vT (m/s) - - - 24.47
For each of the terminal electrostatic velocities, the corresponding

Reynolds number is calculated to ascertain the validity of the simple model. We
can see in Table 2.4, that when the electric field varies from 1 to roughly 104 V/m,
the Reynolds number is well below 1 and this gives us an indication that the
calculated terminal electrostatic velocities are valid approximations.
However, when the electric field is in the order of 106 V/m, the Reynolds
number is above 1 and that means the Stokes approximation does not hold
anymore. It is said that the flow process takes places in the Newton region. Using
equations (2-17) through (2-20), we can correctly calculate the terminal
electrostatic velocity.
Furthermore, from Table 2.4, we can conclude the following. Firstly, when
the electric field stays below roughly 104 V/m, the terminal electrostatic velocity is
well below 1 m/s and this implies that any type of wind that is more than a light
breeze (vw > 3.3 m/s) should be capable of transporting this 1 μm charged droplet
35
away from the charging system. Secondly, however, as soon as the electric field is
in the order of 106 V/m, the terminal electrostatic velocity of the droplet increases,
even after correction for the Newton region, to approximately 25 m/s. For a 1 μm
droplet to be transported away, a storm would be needed (vw > 25 m/s).
The terminal electrostatic velocities for the other droplet diameters listed in
Table 2.3 have been calculated as well. This has been plotted in Figure 2.9, where
the black dots indicate that at the specified electric field the Reynolds number is
equal to 1. At higher electric fields, the Stokes approximation is not valid. When
the droplet diameter is increased to 10 or 100 μm, the Newton region is entered at a
lower electric field.
When the electric field is lower than 105 V/m, the maximum terminal
electrostatic velocities of the droplets vary between 1 to 15 m/s, which is within the
transporting capabilities of normal everyday occurring wind. As soon as the
electric field is in the order of 106 V/m, the terminal electrostatic velocities vary
from 20 to over a 100 m/s. Wind speeds of those magnitudes are much rarer and,
thus, for everyday operation this will pose a problem.
Figure 2.9. The terminal electrostatic velocities of a water droplet as a function of the
electric field. The velocities have been plotted for droplet diameters of 0.1, 1, 10 and
100 μm. The black dots indicate when the Reynolds number is equal to 1.
36
From all this, we can draw a number of conclusions which are useful when
designing an EWICON system. First of all, it is apparent that the electric field
should not be higher than 106 V/m anywhere in the trajectory of the charged
droplets. This means that the electrodes in the charging system should be designed
and graded, such that the charged particles creation process can still take place
without the high fields occurring near the place where the charged droplets are
dispersed. It also means that, when the EWICON system is in operation, its
operating voltage should not rise to a magnitude that the system as a whole
generates an electric field high enough to prevent transportation by the wind.
Another conclusion is that from the range of droplet diameters, for which
we have calculated the terminal electrostatic velocities, most of them are suitable
for charge transportation, depending on the prevalent electric field. If we calculate
the electric mobility of these droplets, we find that it varies from 3·10-5 to 3·10-4
m2/V·s for 0.1 to 100 μm droplets respectively. This partly agrees with the analysis
of Cloupeau [19], who had concluded that for efficient power generation to take
place, the charged droplets need to be created with a low electric mobility in the
order of 10-6 m2/V·s. This value was based on a one-dimensional theory, see [20].
However, in [19] experiments were still conducted with a Type A EWICON
implementation using droplets with mobilities of 3·10-5 m2/V·s. Cloupeau
concluded that wind power conversion with positive efficiency was possible,
stating that actual two-dimensional effects were not taken into account by the one-
dimensional model.
If we were to calculate the electric mobility of droplets for even smaller or

bigger diameters, we would see that the mobility would increase and, therefore,
those droplets would not be suitable for the EWICON system.
2.2.4 Computation & numerical values
Using equations (2-6) through (2-14), a numerical procedure was

constructed which has been entered into MATLAB. For the calculation of the
various quantities such as position, speed and the trajectories of the droplets, a
numerical scheme called Velocity-Verlet has been used. This is a time integration
method that is normally used in molecular dynamics simulations and it is favoured
over other algorithms like Euler integration because of its higher stability and
accuracy or over the Runge-Kutta method because of its relative simplicity.
Without going too much into detail, these methods use a Taylor expansion
of the positions of the particles/droplets as a function of time. In general, based on
the values of the position, velocity and acceleration of these particles and the
potential function at time-step tn, the values of the four quantities can be calculated
at time-step tn+1. After this, the potential function is recalculated using the newly
37
found values for the three quantities and the process repeats itself. The main
difference between the Euler and the Velocity-Verlet method is that the former is a
first order approximation scheme, whereas the latter uses a second order
approximation.
The actual listing of the MATLAB program used for these simulations can
be found in appendix C. A 3D plot of the trajectories based on the numerical
calculations can be seen in Figure 2.10.
Figure 2.10. The resulting plot of the trajectories based on a simulation of three
positively charged droplets in an electric field. The wind speed is set at 10 m/s in the
positive x-direction, the electric field is set at 5·104 V/m in the negative x-direction.
In this simulation, we have set the wind speed at 10 m/s in the positive x-
direction and the electric field at 50000 V/m in the negative x-direction. The
droplets have a diameter of 10 μm and they are positively charged. The charge is
set to 70% of the maximum given by the Rayleigh limit. The simulated time
depicted in Figure 2.10 is 1 seconds.
We can see in Figure 2.10 that the wind is able to overcome the electric
field and, thus, the droplets are transported in the direction of the wind, while
falling down due to gravity. At this point, we look back at section 2.2.1, where we
discussed the evaporation times of charged droplets. For a droplet with diameter of
10 μm, the evaporation time is 0.08 seconds, which means that in this simulation
the droplet would have travelled 0.8 metres. In an EWICON system, this would
38
mean that extra measures would have to be taken before the charged droplets can
reach earth. It can be calculated, however, that in the mist of charged droplets, the
relative humidity is very high and this can prolong the evaporation time of these
droplets up to 10 seconds, see appendix B and/or page 299 of [18].
In this simulation, the droplets have been placed far from each other and,
therefore, the mutual electric repulsive force is not noticeable in the trajectories of
the droplets. This is not the case in a simulation, of which the plot can be seen in
Figure 2.11, where we have put the charged droplets very close to each other and,
consequently, we can see that the repelling electric forces start to have a noticeable
effect on the trajectories of the droplets while they are being transported by the
wind.
Figure 2.11. Plot of the simulation in which ten droplets of equal size and charge are
placed closely to each other. Now, the mutual repelling forces start to affect the paths
of the charged droplets. In the upper left-hand corner (a), an isometric view of the
situation is given; in the upper right-hand side (b) a side view along the direction of
the wind given. In the lower left-hand corner (c) a top view is given; in the lower right-
hand corner (d) a side view in the direction of the wind is given.
39
Together with the plots of the trajectories of the charged droplets, the work
on the droplets performed by the wind on the charged droplets according to (2-14)
has been calculated and it was found to be 1.13·10-8 J. If we assume that a spraying
nozzle disperses roughly 1·107 droplets per second or 20 ml/hr, which, we shall see
in the next two chapters, is not unreasonable to assume, then the power associated
with this stream of droplets is 113 mW per nozzle. We will see in later chapters,
that this is a reasonable estimation for the power delivered by a nozzle capable of
producing these types of charged droplets.
We will take a closer look at scaled up EWICON systems in chapter 5, but

for now we can imagine that if we were to pack 900, e.g. a 30x30 array, of these
nozzles in a system with a wind surface are of 1 m2, the power associated with such
a system would be 102 W.
2.3 Performance Index of the EWICON system
In order to compare the various implementations of the EWICON with the

existing wind turbine technology and with each other, we have to clearly define
numbers that characterise the system performance. This number, which will be
called the EWICON Performance Index (EPI), should include the efficiency ratio,
ηEWICON, of the converted output power, Pout, and the sum of the input power, Σ Pin,
Pout
η EWICON = (2-21)
∑ Pin
where Σ Pin consists of the maximum recoverable power in the wind, the
electrical power needed to charge the droplets and mechanical power to pump the
liquid to the desired height with the desired flow rate.
∑P in = Pmax + Pelectrical + Pmechanical (2-22)
However, (2-21) does not completely represent the nature of the EWICON
system. First of all, since the maximum output power of the EWICON system
depends on the wind surface area, A, it stands to reason that the performance of a
particular implementation is rated higher when it can convert more power from the
same wind surface area.
Furthermore, since liquid droplets are used as charge carriers, the liquid
flow rate, Q, also is an important parameter, which needs to be factored into the
EPI. Lastly, in the case that the efficiency ratio, the liquid flow rate and the wind
40
surface are equal, we want to rate an implementation as higher when it is able to

produce a higher output power.
All these considerations result in an EPI as follows
⎛ η ⎞
EPI ≡ log⎜⎜ Pout ⋅ EWICON ⎟
⎟ (2-23)
⎝ A⋅Q ⎠
where we have taken the logarithm to manage the otherwise high values of
the EPI, which could occur if Pout reaches the order of current conventional wind
turbines. Some example values of the EPI can be found in Table 2.5.
Table 2.5. Example values of the EWICON Performance Index. These four examples
could be characterised as follows: the first two examples can be considered small
prototypes, whereas the last two examples can be considered scaled-up versions. The
last example is rated higher because it requires ten times less liquid to operate at the
same output power.
Σ Pin Q A Pout ηEWICON EPI
(W) (l/h) (m2) (W) (%)
10 1 0.1 1 10 2
100 1 0.1 10 10 3
10000 10 1 1000 10 3
10000 1 1 1000 10 4
2.4 Conclusions
We can conclude that from a theoretical point of view, the EWICON

principle can be used to convert wind energy to electrical energy. The total energy
that can be converted depends on the wind speed, the number of droplets, the
amount of charge placed on these droplets and the strength of the electric field.
Simulations have shown that the work performed by the wind on the droplets
ultimately can lead to powers that are in the same order of magnitude as the power
associated with the wind.
We have found that for a range of droplet diameters (0.1-100 μm), the
electric field should not be higher than 104-106 V/m, if realistic wind speeds (up to
6 on the Beaufort scale) are expected to move the charged droplets away.
41
Chapter 3 – Charged droplet creation methods
3
Charged droplet creation methods
“The noblest of elements is water.”
– Pindar, 476 B.C.
In order for the EWICON system to convert wind energy to electrical

energy, the wind must be able to move charged carriers against the direction of the
electric force on the carriers exerted by the electric field. In theory, these carriers
can be any type of object that can be electrically charged. In the previous chapter,
however, a number of requirements for these carriers, like charge and size, have
been discussed and this means there are certain properties that the carriers ideally
should have.
One of these assumed properties is that these objects should be droplets

and this means that we need methods to create charged droplets from a liquid. For
the sake of completeness, in this research, several experiments have been
conducted using other types of charge carriers, like solid spheres. Without going
into detail, these experiments have proven that the EWICON concept works with
arbitrary types of charge carriers, but these experiments have also proven that most
of these charge carriers are not ideal in terms of ease of creation, level of charging
and power yield. Moreover, most of these non-liquid charge carriers would require
some form of recycling procedure to reuse them. Therefore, the methods described
in this chapter will all focus on the spraying and charging of liquid droplets.
Firstly, in section 3.1, we will look at the other requirements for the
methods posed by the EWICON system, the first one still being that the medium is
a liquid. Subsequently, in section 3.2, we will look at a number of charged droplet
creation methods that have previously been considered and we will discuss the
reasons why these methods have not been implemented in the EWICON system.
In sections 3.3 and 3.4, we will discuss the two main contending charged
droplet creation methods. For each of these two methods, we will discuss the main
working principles, some of the conducted experiments and the results thereof and
43
we will end with the conclusions with respect to their usability within the
EWICON system.
Finally, we will discuss which of the methods are, in principle, most

suitable for the EWICON system, where we should keep in mind that more than
one method is possible, e.g. if the environmental conditions are different.
3.1 EWICON requirements on droplet creation
The efficient generation of charged droplets is probably the single most

important aspect of the EWICON system. Therefore, the choice of the spraying
method should be handled with extra consideration. The chosen spraying
method(s) should at least try to adhere to the following set of requirements, which
will then be explained per point in the following sections:
• low energy consumption with respect to the energy in the wind

• controllable charge, high current output
• monodispersity of the liquid droplets
• operable under windy condition
• environmentally friendly spraying liquid
3.1.1 Low energy consumption with respect to the energy

in the wind
There are several methods available for creating charged droplets. Usually,
these methods require high electric fields and/or high pressure to operate. Some of
these methods require significant amounts of energy and this means that it would
be more difficult to achieve positive conversion efficiency. Of course, the usability
of a method depends on the prevailing wind speed at the site location and its
distribution and on how much energy other components in the EWICON system
need.
For example, if we take wind through a surface area of 1 m2 with a speed

of 10 m/s, then we know, by using equation (2-1), that the power associated with
this particular wind is 647 W. Knowing, with Betz in the back of our mind, that
roughly 60% of the power in the wind can be converted into electrical energy, this
means that we can extract 383 W at most. So, if a particular spraying method
would need e.g. 221 W to operate leaving 162 W as the net output, the maximum
conversion efficiency would be 25%. This does not even include other factors like
e.g. monitoring equipment or supplying the liquid to a certain height.
44
3.1.2 Controllable charge, high current output
In the previous chapter, calculations have been performed with respect to

the electrical mobility of the charged droplets and the considerations made by
Cloupeau 0. It was concluded that for a certain range of mobilities the performance
of the EWICON system would be optimal. Therefore, the possibility to control the
charge, that is present on a droplet, would give us a possibility to fine-tune the
EWICON system such, that all the charged droplets would be transported away.
Also, in the case that it would be relatively simple to adjust the amount of charge,
the EWICON system could be designed in such a way that it could easily adjust to
changing wind speeds.
The objective of the EWICON system is to convert as much wind power to

electrical power as possible. This means that the current associated with the
charged droplets should be as high as possible, assuming that the EWICON system
operating potential is constant. This not only requires the droplets to be highly
charged, but also that they are produced at a high rate.
In order to determine what is considered a high current, we will consider a

small example. Using the information of chapter 2 as a guiding principle, we look
at water droplets with a diameter of 5 μm that are sprayed from one nozzle. If these
water droplets can be charged to 70% of the maximum charge given by the
Rayleigh limit, then the current per nozzle can be calculated based on the flow rate
of the liquid. At 20 ml/hr, the rate of charged droplets is 8.5·107 per second and this
rate amounts to a current through the nozzle of 4.7 μA per nozzle. If we have an
electrical load of 20 GΩ, then this would imply an output power of roughly 0.5 W.
If we revisit the situation described in section 3.1.1, in which we wanted to

achieve a conversion efficiency of 25%, then this would imply that we require 212
nozzles per m2 to produce the current required to feed the load. Now, if we take
droplets with a diameter of 15 μm, then the current drops to 0.9 μA per nozzle,
which implies an output power of 17 mW. This would come down to roughly 5500
spraying nozzles per m2. In terms of liquid consumption, this equals going from 4.2
litres per hour for the former situation to 110 litres per hour for the latter situation.
3.1.3 Monodispersity
The production of monodisperse droplets means that droplets of only one

size are created. Together with the requirement described in the previous section,
we would have a droplet with one particular size and charge. This is advantageous
for the EWICON system for a number of reasons. First of all, as we have seen in
the previous chapter, the operation of the EWICON system is greatly simplified
45
regarding control purposes by only having to take into account one particular type
of droplet.
Secondly, in monodisperse mode, the EWICON behaves like a near-ideal

current generator, i.e. it can deliver a constant current to a load for any working
voltage up to a certain maximum voltage. The fact that all droplets are created with
equal size implies that they all have the same upper charge limit related to the
Rayleigh break-up principle. This means that the droplets are all charged in the
same manner, resulting in droplets with equal charge. In turn, this means that, on
average, the force balance on each of the charged droplets, as described in the
previous chapter, would be the same.
Therefore, if one charged droplet can be moved against the electric field by
the wind, then all the droplets could be moved. Conversely, if one charged droplet
is attracted back to the EWICON system, because e.g. the wind speed has dropped
or the EWICON system potential has increased, all the droplets would be attracted.
And thus, the output current would drop to zero at that potential. This does not take
into account space charge effects, which causes charged droplets closer to the
EWICON system to be pushed back by droplets that have been dispersed earlier.
In practice, there will always be a range of droplet sizes and, therefore,

monodispersity as explained in the first paragraph will usually not occur. In aerosol
science, a droplet distribution is defined as monodisperse when the geometric
standard deviation σg or GSD of the droplet size distribution is smaller than 1.2, see
e.g. [21],
σ g < 1.2 (3-1)
in which σg is given by
⎛ ⎞
∑ (ln d )
n
⎜ i − ln μ g
2
⎟
⎜ ⎟
σ g = GSD = exp σ = exp⎜ i =1
⎟ (3-2)
n
⎜ ⎟
⎜ ⎟
⎝ ⎠
in which n is the number of measured droplets with a diameter di, σ is the

standard deviation of the log-normal distribution and μg is the geometric mean
defined by
46
μ g = n d1d 2 ...d n (3-3)
This assumes that the distribution of the produced droplets follows a log-
normal distribution. For the measured droplets in the subsequent parts of this
thesis, this will be determined by using statistical methods.
For large groups of droplets, calculating the GSD by using (3-2) and (3-3)
can become laborious and, therefore, the GSD can also be estimated by
d 84
GSD = (3-4)
d16
in which d16 and d84 are the droplet diameters at the 16% and at the 84%
size cut of the cumulative distribution. Equation (3-4) will be used, when the
statistical fit of a distribution cannot be accurately determined.
After performing droplet size measurements, which will be discussed in

this chapter and in chapter 4, the monodispersity of the produced droplets can be
determined by using one of these methods.
3.1.4 Operable under windy condition
As we will see, there are a number of charging methods that provide a high
current output and monodispersity in a wind-free environment which would suit
the EWICON system, but these methods fail to deliver the same results when a
flow of air is applied. Thus, for each method under consideration, it should be
tested that the creation of charged droplets is not disturbed by the wind.
3.1.5 Environmentally friendly spraying liquid
The EWICON system is an open system, which means that it will be in

contact with the surrounding environment, dispersing its charged droplets into it.
Therefore, the spraying liquid needs to be safe for the environment if the EWICON
is to be accepted as an alternative for the wind turbine. Preferably, the used
charging method(s) should be able to make use of water, saline or fresh.
47
3.2 Previously considered creation and charging

methods
There are several spraying and/or charging methods available and a

number of these methods have been considered and tested in a previous EWICON
research project, see [22]. We will list two droplet creation methods and one
droplet charging method, give a brief description and explain why these methods
have not been implemented in the current EWICON system.
3.2.1 Droplet creation - Ultrasonic atomisation
Ultrasonic atomisation is a method, with which droplets can be produced

by using ultrasonic waves at the liquid-air interface, see e.g. [23]. These vibration
forces, responsible for the atomisation, are generated by piezoelectric ceramic
crystals. Two mechanisms are believed to be responsible for the droplet break-up,
namely the capillary wave hypothesis and the cavitation hypothesis. The reader is
referred to the references for these mechanisms. Droplets generated by this
atomisation method have relatively small diameters, ranging from a few microns to
a few tens of microns. Also, the distribution of the droplet size can be made
narrow.
The main drawback of this atomisation method is that it is quite costly in

terms of power dissipation. Devices capable of ultrasonic atomisation require 1 to
20 kilowatts to operate depending on the required flow rate. This would mean that
an EWICON system would require a wind surface area of 2 to 50 m2 operating at
the Betz optimum conversion efficiency just to provide power for the droplet
generation.
3.2.2 Droplet creation - Centrifugal or rotating disk

atomisation
If a liquid is fed to the centre of a rotating disk in the form of a continuous

jet at a small flow rate, then this liquid can be atomised, see [24]. Such a rotating
disk can produce small droplets at high rotating speeds. Also, this technique can be
modified to an electrified rotating disk in order generate electrospray.
The main drawback of this atomisation method is the fact that it is based on
rotational movement and, hence, susceptible to wear and tear. This, of course, is
the very reason why the concept of the EWICON system was introduced in the first
place. Consequently, maintenance costs (together with purchase costs) of this type
of atomiser are relatively high. Other drawbacks include the lack of monodispersity
of the droplets.
48
3.2.3 Droplet charging - Corona charging
Both methods that are described in 3.2.1 and 3.2.2, in principle, only
provide uncharged droplets (although for rotating disk a charging method has
already been given). Therefore, they have to be combined with a charging method
to obtain charged droplets.
Corona is a partial breakdown effect that occurs when the local electric
field surrounding a sharp point, e.g. a charging electrode, becomes too high. When
this happens, ionisation occurs because electrons are accelerated away (in case of
negative potential) from the electrode, colliding with nitrogen or oxygen
molecules, see [25]. This results in free electrons, which, in turn, can also collide
with other molecules, leading to an electron avalanche. Typically, electrical
breakdown occurs at electric fields of 3.0 kV/mm in air, but this also depends on
the radius of the sharp point. In general, corona is an unwanted effect because it
causes e.g. losses in high voltage lines or radio interference.
However, corona can also be used for the charging of particles and it is
widely used in the coating industry or for the precipitation of dust. Therefore, this
charging method has been considered for the application of charge on the droplets,
which could be created by e.g. one of the two previously mentioned methods. The
ions that have been created by the corona process will attach themselves to the
droplets, a process called “ion bombardment”.
The power required to perform corona charging, Pcorona, is given by
Pcorona = Vcorona ⋅ I ion (3-5)
in which Vcorona is the potential of the charging electrode at which the

corona discharge occurs and Iion is the current associated with the ionisation.
Typical inception potentials for corona are in the range of 450 V to 15 kV per
electrode for air depending on the polarity of the potential and the radius of the
sharp edge of the electrode, see [25]. Currents associated with these potential are in
the range of 200 to 1000 μA, see [26]. Consequently, typical power requirements
vary from 0.1 to 15 W per electrode.
One of the drawbacks of this charging method is that the created droplets
need to be directed along the area where the corona is created for them to acquire
charge. Inevitably, there will be a number of droplets that will not be charged. This,
in turn, means that the current associated with the charged droplets will always be
smaller than the current associated with the ionisation.
49
3.3 Electrohydrodynamic atomisation
Electrohydrodynamic atomisation, EHDA or electrospray, is a spraying

and charging method which is e.g. used for coating purposes or medicine
administering. Its main advantage is that very little energy is required for the
creation of charged droplets.
3.3.1 EHDA principle
The EHDA spraying method is based on the principle that a strong electric
field will deform the meniscus of the liquid leaving a spraying nozzle to a conical
shape, i.e. the electric Coulomb force will interact with the surface tension of the
liquid. Ions in the liquid will accelerate towards the cone apex and, due to this
movement, the liquid itself is accelerated. At the cone apex, a liquid jet occurs
which breaks up into droplets with a high charge density. Depending on the
direction of the applied electric field, the net charge on the droplets will either be
positive or negative. Usually, this electric field will be created using charging
electrodes.
The droplet diameter, charge and initial velocity as well as the rate at
which the charged droplets are ejected from the nozzle, all depend on the potential
of the charging electrodes, the geometrical configuration of these electrodes and
the properties and the flow rate of the spraying liquid.
Using EHDA, it is possible to spray droplets in different spraying modes,

depending on the flow rate of the liquid and the shape and strength of the local
electric field. The spraying modes of EHDA are separated into two general
categories: modes that exhibit a continuous flow of liquid through the meniscus
and modes that do not. The former consists of the simple-jet, the cone-jet and the
ramified-jet, while the latter consists of the dripping, the micro dripping, the
spindle and the intermittent cone-jet modes. The latter are often referred to as
pulsating modes, see [27].
One of these continuous spraying modes is the cone-jet mode or Taylor

cone mode, see Figure 3.1, and one of its main advantages is the fact that droplets
are highly charged up to 70% of the maximum charge given by the Rayleigh limit,
see [28]. Also, the droplet distribution in this cone-jet mode can be monodisperse.
50
Figure 3.1. An example of EHDA spraying in the cone-jet mode, which is also referred
to as the Taylor cone mode. Ethanol is sprayed under the influence of an electric field
which results in a stream of charged droplets. The droplets all have the same polarity,
which causes the expanding mist of droplets.
One drawback of EHDA, however, is that in order to spray in the cone-jet

mode, the surface tension of the used spraying liquid should be relatively low.
Using a conventional EHDA spraying set-up with demineralised water, continuous
cone-jet spraying does not occur due to the high surface tension of water (72·10-3
N/m); see [29]. In order to actually achieve any spraying of charged water droplets,
the electric field has to be increased to such a level, that it leads to a very unstable
spray without monodisperse droplets. Part of this instability is caused by the fact
that the charged water droplets reduce the electric field at the nozzle, thereby
cancelling the cone-jet spraying as it has started. These charge relaxation times are
usually in the order of 100 ms. This does not occur with ethanol, because, firstly,
the electric field is not required to be as high as with water and secondly, the
charged ethanol droplets have less charge on it than charged water droplets.
Also, in order to achieve such a high electric field, the charging electrodes
have to be set on such a high potential that corona discharges start to occur.
Ethanol e.g., on the other hand, has a much lower surface tension (22·10-3 N/m) and
spraying in the cone-jet mode is easily achieved.
Liquid conductivity also plays a role in the feasibility of spraying in the

cone-jet mode. According to Cloupeau [30], the flow rate at which cone-jet mode
can be achieved decreases with increasing conductivity. A lower flow rate,
however, could mean that a smaller number of charged droplets would be produced
per time unit. Therefore, the advantage of spraying in cone-jet mode, i.e. higher
charge per droplet, should outweigh the lower rate of produced droplets.
In practice, when using water, electrospraying will occur in one of the

other aforementioned spraying modes of EHDA, each with their own
characteristics. However, none of these spraying modes spray monodisperse
droplets, meaning that we have to take into account that the droplet diameters will
51
follow a distribution. Also, none of these modes produce droplets that are charged
as high as the droplets in the cone-jet mode.
3.3.2 Analytical model: EHDA
Several models exist for describing the EHDA spraying process. Using
these models, it is possible to give an estimation of the droplet size and current
output produced by the charged droplets as a function of the flow rate and/or the
applied electric field. The model used in this research has been developed for the
previously mentioned cone-jet mode and, therefore, it is only indicative for the
other modes of spraying. Also, this model does not take directly into account the
electric field. It assumes that the electric field has enabled cone-jet spraying and
then it calculates the droplet size and charge.
The model, however, still can be used to gain a better understanding about
how parameters like flow rate, surface tension, conductivity and
(absolute/dynamic) viscosity affect the spraying process. We will not cover this
model in detail, but we will state the important conclusions needed to effectively
analyse and, ultimately, use EHDA spraying for the EWICON system.
Gañán-Calvo [31] found that the relations for the droplet size and current
differ significantly depending on the aforementioned parameters, especially
between highly conductive viscous liquids and liquids with a low viscosity and
conductivity. Therefore, a viscosity number, VN, was introduced to determine
which relation was valid for a given spraying liquid in the cone-jet mode,
1
⎛ γ ε ⎞
3 2 3
VN = ⎜⎜ 3 ⎟⎟ 0
(3-6)
⎝μ K Q⎠
2
in which μ is the dynamic (or absolute) viscosity, K is the conductivity and

Q is the flow rate. From VN we can see that for a specific liquid, the relation
between the flow rate and the conductivity is given. The higher the conductivity of
the liquid, the lower the flow rate should be to keep the viscosity number, VN,
constant.
In the case of low viscosity and low conductivity, for liquids such as water
and ethanol, the viscosity number, VN, likely will be larger than 1. In that case, the
current and droplet size can be estimated using the following two relations [31],
52
1 1
dd ⎛Q ⎞2 I ⎛Q ⎞4
= 1.2⎜⎜ ⎟⎟ − 0.3 , = 11.0⎜⎜ ⎟⎟ − 5.0 (3-7)
do ⎝ Qo ⎠ Io ⎝ Qo ⎠
in which dd is the droplet diameter and Io, do and Qo are the characteristic
current, droplet diameter and flow rate of the spraying liquid, given by the
following set of equations
1 1
⎛ε γ 2 ⎞2 ⎛ ε 2γ ⎞3 ε 0γ
I o = ⎜⎜ 0 ⎟⎟ , d o = ⎜⎜ 0 2 ⎟⎟ , Qo = (3-8)
⎝ ρ ⎠ ⎝ ρK ⎠ Kρ
in which ρ is the density of the liquid. If we consider demineralised water

with a conductivity of 2.0 μS/cm, then we can use equations (3-8) to calculate the
characteristic current, the droplet diameter and the flow rate as shown in Table 3.1.
Table 3.1. Calculation of the characteristic current, Io, droplet diameter, do, and flow
rate, Qo, of demineralised water with a conductivity, K, of 2.0 μS/cm, a surface tension
of 72·10-3 N/m and a density of 998 kg/m3.
Io 6.86·10-3 μA
do 5.23·10-1 μm
Qo 1.16·10-2 ml/hr
Using these three values together with equations (3-7), we can get
estimations for the droplet diameter, dd, and the current, I, associated with a droplet
creation process in the cone-jet mode. For example, if we set the liquid flow rate,
Q, at 1 ml/hr, then I and dd become as shown in Table 3.2.
Table 3.2. The estimated current, I, the droplet diameter, dd, and the Rayleigh
percentage, q/qmax, when using EHDA spraying with demineralised water in the cone-
jet mode for example flow rates, Q, of 1 and 20 ml/hr.
Q = 1 ml/hr Q = 20 ml/hr
I 0.20 μA 0.45 μA
dd 5.66 μm 25.87 μm
q/qmax 70 % 78 %
We now have an indication of the magnitude of the currents per nozzle of

an EWICON system that is fitted with an EHDA spraying system. We have also
included the charge, q, on the droplet as a percentage of the maximum given by the
53
Rayleigh limit, qmax. As expected, this model predicts that when EHDA spraying is
performed in the cone-jet mode, the droplets are highly charged.
Secondly, if we calculate VN using a flow rate of 1 ml/hr, then we find that

it yields a value of 1.6, which is greater than 1 and, thus, we can legitimately use
the relations stated in (3-7). However, if we increase the flow rate to, for example,
20 ml/hr, then we find that VN yields a value of 0.6, meaning that we have to use
different relations, again from [31]:
1
dd
1
⎛Q ⎞3 1
= 1.6(ε r − 1) 6 ⎜⎜ ⎟⎟ − 1.0(ε r − 1) 3 ,
do ⎝ Qo ⎠
1
I 6 .2 ⎛ Q ⎞2
= ⎜ ⎟⎟ − 2.0 (3-9)
I o (ε − 1) 14 ⎜⎝ Qo ⎠
r
in which εr is the relative liquid permittivity, which in the case of water is

80.1 at 20°C. In Table 3.3, we have calculated I and dd for a flow rate, Q, of 20
ml/hr. For comparison, in Table 3.2, we have also included I and dd for the same
flow rate using equations (3-7).
Table 3.3. The estimated current, I, the droplet diameter, dd, and the Rayleigh
percentage, q/qmax, when using EHDA spraying with demineralised water in the cone-
jet mode for example flow rates, Q, of 20, 40 and 80 ml/hr. This time, the viscosity
number has decreased to 0.6 or less, thus the calculations have been performed using
equations (3-9) instead of equations (3-7).
Q = 20 ml/hr Q = 40 ml/hr Q = 80 ml/hr
I 0.58 μA 0.82 μA 1.17 μA
dd 18.53 μm 23.92 μm 30.73 μm
q/qmax 61 % 64 % 66 %
Finally, another observation, which can be made from Table 3.3, is that
when the liquid flow rate is increased, the current follows, but not proportionally.
As we can see, the flow rate needs to be quadrupled to double the current.
Apparently, increasing the flow rate in the cone-jet mode causes the droplet
diameter to increase and this means that the charge to mass ratio decreases. This is
an important effect linked with some of the spraying modes produced with EHDA,
which should be kept in mind when the electric current production needs to be
increased in order to make use of higher wind speeds.
54
The extra wind power conversion that could be gained from the current
increase should thus be weighed against the disadvantages of supplying liquid at a
higher flow rate; which are amongst other things more dispersed liquid and more
pumping power. We will discuss this in more detail in chapter 4 using the
EWICON Performance Index as defined in chapter 2.
3.3.3 EHDA implementation and experiments
The goal of the following experiments was to determine how to actually

establish EHDA spraying. We will look at the basic principles of EHDA in practice
and analyse and design a system that will be able to electrospray various liquids.
We will, then, measure the produced current associated with the spraying process,
which will give us an indication of its potential use in the EWICON system. The
question of whether the following EHDA spraying methods will actually perform
as expected when there is wind present, will be discussed in chapters 4 and 5,
where the overall EWICON design and the accompanying experiments are treated.
The principle in practice
In principle, a charging system based on droplet creation should consist of

one or more spraying nozzles, a reservoir for the spraying liquid and electrodes to
generate the electric field to establish the spraying process. These principles also
apply to an EHDA based spraying system and, here, we will first look at a
configuration with a single nozzle and a single electrode, schematically depicted in
Figure 3.2.
55
Figure 3.2. Schematic depiction of an EHDA based spraying system. The liquid is
supplied through a stainless steel needle nozzle, while the electric field is generated by
a copper ring electrode connected to high voltage source. In this case, the inner
diameter of the ring was 2.0 cm and the needle was placed in the centre.
An actual implementation of such a configuration, called the “Delrin”

nozzle because of its construction material, can be found in Figure 3.3. A stainless
needle nozzle connected to a syringe pump has been used to supply liquid to the
system. The testing liquids were demineralised water, ethanol and various mixtures
of both, because mixing ethanol with water lowers the surface tension of the liquid.
Flow rates have been tested varying from 1 to 40 ml/hr. A copper ring
acting as the charging electrode positioned around the nozzle was connected to a
high voltage DC source. The potential on the electrode was varied from ±1.0 to
±7.0 kV. The spraying needle is connected to earth potential and this means that
the electric field schematically will resemble the situation depicted in Figure 3.4a
and Figure 3.4b. The charged droplets were sprayed towards a metal plate that was
connected to ground.
56
Figure 3.3. One of the possible implementations of an EHDA spraying system, called
the “Delrin” nozzle. Two needle nozzles of different lengths are displayed on the right.
The liquid is fed through connection on the upper side of the device. The ring
electrode and the connection to the HVDC source are shown on the lower side.
Figure 3.4. The approximate direction of the electric field in an EHDA spraying
system. In both case, the spraying needle is connected to earth. In a) the charging
potential is positive, so the electric field points inwards. In b) the charging potential is
negative and the electric field points outwards.
Electrospray was achieved using this simple configuration, but stable

spraying, however, was only observed when pure ethanol or water/ethanol mixtures
were used as the spraying liquid. Cone-jet mode was only achieved with ethanol.
57
Since no wind was present during this stage of the testing, all the created charged
droplets were instantly attracted back to the charging electrode.
Visual inspection with laser light of the spraying process resulted in Figure
3.5, in which, on the left-hand side, the process is shown for a positive charging
potential. On the right-hand side, we can see the process for a negative charging
potential.
Figure 3.5. A photo of the spraying process using water/ethanol as the spraying liquid.
Both a positive charging potential, on the left, and a negative charging potential, on
the right, was used. The spray is illuminated with laser light. Note the presence of a
fine mist on the right.
For both negative and positive charging potentials, currents were measured
up to 0.5 μA with a 70%/30% water/ethanol mixture per nozzle. The higher
currents were obtained using a flow rate of 20 ml/hr. Higher flow rates than 20
ml/hr did not significantly increase the currents. When demineralised water was
used, the currents were less than 0.1 μA.
In order to get an indication of the electric field needed to establish

electrospraying, a simulation software package called LORENTZ has been used to
calculate the electric field generated by the ring electrode as shown in Figure 3.2,
see appendix E. On average, electrospraying commenced when the ring electrode
was on a potential of -4.0 kV. The electric field 1.0 mm underneath the nozzle tip
was found to be 7.75·105 V/m. At -6.0 kV, where the electric field is 1.16·106 V/m,
the current levelled off.
During these experiments, the current flowing from the charging electrodes
to the high voltage DC source was also measured. This was found to be lower than
0.05 μA for a charging potential of -6.0 kV and, therefore, the power dissipation
associated with the EHDA spraying process is estimated to be in the order of a few
nanowatts per nozzle.
On the right-hand side of Figure 3.5, we can also see the presence of a fine
mist. Since this mist apparently did not affect the produced current, we will not
discuss it in this section. Instead, we defer the discussion of this mist and its effect
on droplet transportation in chapter 4, when wind is introduced into this EHDA
system.
58
Measuring the size, velocities and charge of the droplets
Using a Laser Phase-Doppler analyser (LPDA), we were able to determine

the droplet diameter and velocities of the EHDA spraying system. In Figure 3.6,
the droplet diameter distribution is given for the Delrin single nozzle system
spraying demineralised water using a flow rate of 20 ml/hr.
For the measurement data collected by the LPDA the geometric mean, μg,
and the standard deviation, σ were calculated. According to (3-2), the geometric
standard deviation, GSD or σg, can be derived by taking the exponential of the
standard deviation. For this particular measurement, the GSD was calculated to be
1.6, which means that the droplet distribution is not considered monodisperse.
Figure 3.6. The droplet diameter distribution as measured by a Laser Phase-Doppler

set-up with water. The calculated GSD of this distribtion is 1.6. However, it is more
likely to be higher, because very small and large droplets have not been registered by
the LPDA.
This does not paint the whole picture, though, because not all droplet sizes
have been measured. As previously stated, electrospraying water by conventional
59
methods is a very chaotic process, and therefore, a significant part of the droplets,
smaller than 10 μm and larger than 100 μm, are not registered by the measuring
volume of the LPDA. If these droplets are also taken into account, the GSD of this
distribution is likely to be higher than 2-2.5.

set-up with water/ethanol. The GSD of this distribution is 2.0.
In Figure 3.7, the droplet diameter distribution is given for the Delrin
single nozzle system spraying demineralised water mixed with ethanol using a flow
rate of 20 ml/hr. Now, the GSD was calculated to be 2.0. Again, this means that the
droplet is not considered monodisperse.
During these same experiments, the velocities were also measured in the
directions as specified in Figure 3.8. The velocities in the x-direction were found to
be varying between 1 and -1 m/s. In the z-direction, the velocities were found to be
varying between 1 and 5 m/s.
60
Figure 3.8. The velocities of the droplets have been in measured in the z and x
direction.
Also, by using a system called ELPI (Electrical Low Pressure Impactor),

we were able to obtain a rough indication of the charging efficiencies on the
droplets produced with EHDA spraying and whether these efficiencies varied as a
function of the droplet diameter. In the ELPI, the droplet are collected in different
impactor stages according to their diameter and the electric charge carried by the
droplets into each impactor stage is measured in real time by sensitive
electrometers. For more information, see [32].
These measurements indicated that the water droplets created with EHDA
were charged up to 30% of the maximum charge as given by the Rayleigh limit. In
section 3.3.1, we stated that the electric field strength is reduced by the creation of
the charged droplets, thereby preventing the water spraying process to enter the
cone-jet mode. This reduced electric field is also the reason that the water droplets
are not as highly charged as expected with the cone-jet mode.
Affecting the surface tension reduction of the liquid by CO2
Earlier, the surface tension of demineralised water was lowered by adding

ethanol. However, this is not a long term solution if the EWICON system is to be
used for large scale commercial use, both from an environmental and financial
point of view. Therefore, other methods have been devised to lower the surface
tension of water. One of these methods involves the use of CO2, since the surface
tension depends on the two media between which the surface is formed.
61
In [33], the surface tension of water was lowered from 72·10-3 N/m to
57·10-3 N/m as CO2 was added to water. This implies that the electrospraying of
water could be improved by using gaseous water. Experiments have been
conducted, of which the results will be briefly given.
Carbonised water with a conductivity of 105 µS/cm has been sprayed in a
Delrin nozzle system using similar parameters as used with demineralised water
and the water/ethanol mixtures. This yielded a current of 0.1-0.2 μA per nozzle,
whereas spraying pure demineralised water yielded less than 0.1 μA per nozzle.
Using carbonised water in this particular EHDA system gives significantly better
results compared to demineralised water in terms of produced net current.
However, technical difficulties arose while the carbonised water was fed to
the spraying nozzle. In the water reservoir, the CO2 escaped from the water causing
extra pressure that increased the flow rate of the liquid. This led to fluctuations in
the produced current. On occasion, large CO2 bubbles entered the tubing system,
which led to interruption of the liquid flow. This meant that no charged droplets
could be produced and the current was zero.
Cylindrical electrode/nozzle configuration
When it became apparent that the stable electrospraying of (demineralised)

water was not possible with the previously discussed configuration, a more detailed
analysis was made in LORENTZ of the potentials and electric field distribution in
the existing system. In Figure 3.9, results of the equipotential analysis are shown
for the standard Delrin electrode configuration.
Figure 3.9. Equipotential plot around the tip of the spraying needle of an EHDA set-
up. The charging electrode set at -4.0 kV is not shown in this plot. It was set at -4.0 kV.
The spraying needle was connected to earth potential.
62
In this plot, the charging electrode was set at -4.0 kV and the spraying
needle was set at earth potential. For the sake of clarity, the electrode is not shown
in Figure 3.9. In order to analyse the effect of the electrode, this analysis was
conducted without the liquid present.
If we look at the equipotential lines in the plot shown in Figure 3.9 around
the tip of the needle, the field concentration is limited, due to the distance of the
electrode to the spraying needle tip, which was approximately 2.0 cm in this case.
Hence, in order to achieve some level of stable spraying the potential of the
electrode would need to be increased, but, as previously mentioned, this would lead
to corona discharges.
This knowledge was used to investigate whether spraying water is possible

using a somewhat more complex electrode system. This led to the concept of a
cylindrical electrode, partly based on [34] with a diameter of 4 mm that was
designed around a stainless steel needle nozzle. The advantage of having such a
small diameter was the fact, that the electric field could be focussed on the tip of
the spraying nozzle, while keeping the electrode itself on a relatively low potential.
In this way, corona discharge effects could be kept to a minimum. A photo of this
cylindrical electrode/nozzle configuration can be seen in Figure 3.10. The tip of the
needle sticks out approximately 0.5 mm.
Figure 3.10. A spraying needle with a cylindrical electrode. The cylindrical electrode
has a diameter of 4.0 mm and the spraying needle sticks out 0.5 mm.
Again, an equipotential plot was made of this needle/cylindrical electrode

configuration with the cylindrical electrode set at -2.0 kV and the needle set at
earth potential, see Figure 3.11. In this plot, we can see that the lines are
concentrated around the tip of the needle, ensuring that a high electric field is
present only at the tip of the spraying needle.
63
Figure 3.11. Equipotential plot of the cylindrical electrode EHDA set-up designed to
spray water. Part of the cylindrical nozzle can be seen with the tip of the spraying
needle sticking out. The cylindrical electrode potential was set at -2.0 kV. The
spraying needle was connected to earth potential.
Experiments have been conducted with a configuration as depicted in

Figure 3.10. The first and most important observation was that stable
electrospraying with demineralised water was achieved. Stable spraying in this
context means that droplets were sprayed downwards in a non-intermittent manner.
However, the spraying was not in the cone-jet mode.
For a flow rate of 20 ml/hr, the measured current was 0.3 μA compared to
the less than 0.05 μA found with the Delrin nozzle system.
Using similar spraying parameters as for the Delrin nozzle system, the
droplet diameter distribution has again been determined using the LPDA, of which
the results can be seen in Figure 3.12. This time using (3-4), the GSD was found to
be 1.24, which means the droplet diameter distribution is considered to be nearly
monodisperse. Thus, as far as electrospraying demineralised water is concerned,
this is a significant improvement over electrospraying water with the Delrin nozzle
system, as was shown in Figure 3.6.
64

set-up with demineralised water using the improved spraying system. The GSD of this
distribution is 1.24, which means that the distribuiten is nearly monodisperse.
Stable spraying with tap or saline water, however, was still not possible as
the high conductivities of both types of water caused very chaotic spraying.
Typical values are 5·104 μS/m for tap water and 5·106 μS/m for sea water as
compared to the 200 μS/m for demineralised water. This is mainly due to the fact
that with electrospray high conductivities lead to unstable electric dripping modes.
According to [29], this situation can be partly resolved by decreasing the

flow rate of the tap or saline water to a few millilitres per hour. However, this
would mean that the number of nozzles needs to be increased by a factor of at least
10 to maintain the produced current. Another solution that is given in [29] is to
increase the electric field by increasing the potential on the electrodes. However,
the required potentials would need to in the order of 20 kV. In this cylindrical
electrode configuration, this would inevitably lead to discharges.
65
3.3.4 Conclusions on EHDA based spraying systems
First of all, the experiments have shown that EHDA can create droplets
suitable for an EWICON system in terms of droplet size and charge.
Monodispersity was not present for any tested liquid except pure ethanol. Using a
70%/30% water/ethanol mixture, the size distribution was found to be narrower.
Unfortunately, the degree of monodispersity found with the electrospraying of pure
ethanol was not achieved.
Electrospraying water mixed with CO2 did yield higher currents than
electrospraying with pure demineralised water. However, this did bring along other
practical issues, which leads to the conclusion that, currently, this approach is not a
viable solution for electrospraying water.
Finally, it has not yet been possible to electrospray water in cone-jet mode.
Stable electrospraying yielding a higher current, however, was achieved using a
cylindrical electrode. Also, the droplet size distribution was found to be nearly
monodisperse. This means that, with EHDA, an EWICON system can be designed
that uses water as a spraying medium and, thus, no harmful liquids are dispersed
into the environment.
3.4 High pressure monodisperse spraying1
The second method of charged droplet creation that has been investigated
is the method called high pressure monodisperse spraying (HPMS). This method is
based on the principle that a liquid is forced through a device fitted with small
micron-sized pores with equal size creating liquid jets with equal diameter. The
high pressure that is applied is usually in the order of 10 to 15 MPa. The liquid jets
break up into droplets due to the Rayleigh break-up principle, with the diameter of
the droplet proportional to the diameter of the liquid jet.
This droplet creating method has primarily been developed for continuous
inkjet printing and, as such, was designed to operate with all kinds of liquids,
including saline water. Based on the preceding section covering the EHDA
spraying method, we have concluded that the creation of nearly monodisperse
droplets with water is possible, but requires a more complex spraying system. Only
ethanol will result in actual monodisperse droplets. Therefore, with HPMS, there is
already a significant advantage of this method over the EHDA method, because the
intended spraying liquid is water.
1
This work was conducted in close cooperation with Wietze
Nijdam and Jeroen Wissink of Medspray XMEMS BV
66
3.4.1 HPMS principle
While the droplets are created, the droplets are charged by induction using
charging electrodes. In general, an HPMS set-up will schematically resemble a
configuration like depicted in Figure 3.13. Because the charging process is
separated from the droplet creating process, this means that, in theory, the charged
droplets could be fine-tuned more easily to suit the specific EWICON conditions.
Figure 3.13. The working principle of the HPMS spraying method is depicted. A
liquid is forced through a micron-sized pore resulting in a jet. This jet breaks up into
monodisperse droplets while being charged by electrodes. In this case, the electrode is
on a positive potential resulting in negatively charged droplets.
In this case, the liquid is connected to earth and the electrode(s) is set to a
positive potential thus causing the droplets to be negatively charged. A negative
charging potential, conversely, will yield positively charged droplets. Similar to the
EHDA spraying method, these electrodes can be designed in several ways. The
charging electrode depicted in Figure 3.13 can for example be designed as a ring
electrode as shown in Figure 3.14. This has been analysed in LORENTZ.
67
Figure 3.14. A ring electrode has been modelled for the water droplets created by the
HPMS system. The droplets are charged by induction. The charge on a droplet can be
determined by calculating the capacitance between the ring electrode and the droplet.
Another observation that can be made with respect to the charged droplets
created by HPMS is the exit velocity. Whereas the droplets exiting the spraying
nozzles in an EHDA based spraying system usually have a velocity of 1-5 m/s, the
droplets in the HPMS system have a velocity of 10-20 m/s. Therefore, if we look at
the calculations made in the last chapter on droplet mobility, maximum electric
fields and minimum required wind speeds, we can expect that the charged droplet
can be removed more easily.
3.4.2 Analytical model: HPMS
Firstly, we will calculate and predict the amount of current we can expect
from this spraying system assuming a certain flow rate, a charge per droplet and
droplet size. We will also look into the charging process itself.
These HPMS devices that have been tested in the current EWICON system
were designed to produce droplets with a diameter of 20 μm. For these
experiments, a flow rate of 71 ml/hr per device was used and, this means that
68
theoretically, the number of droplets per second that should be dispersed is 4.7·106.
If we assume that these droplets can be charged up to 30% of the maximum charge
as given by the Rayleigh limit, then the charge per droplet would be 1.9·10-13 C.
Thus, the current associated with the stream of droplets is 0.9 μA.
The charging process of the droplets is similar to the process implemented

in ink jet printer technology. In this technology, monodisperse ink droplets are
firstly charged by charging electrodes and then guided to their place on the printing
paper by deflection electrodes.
The charging of the droplets can be modelled by determining the

capacitance Cdroplet between the droplet jet and the electrode.
q droplet = C droplet ⋅ V (3-10)
in which qdroplet is the charge on the droplet and V is the potential on the
charging electrode. The capacitance can either be obtained by analytical means or
by using a simulation software package as LORENTZ as shown in Figure 3.14.
Firstly, we will give two analytical calculations. The first one is based on a
sphere-plate configuration. The capacitance between a conductive sphere with a
radius ra and a metal plate at a distance d is given by [35]
⎡ 1 ⎤
C droplet = 4πε 0 ra ⎢1 + ln (1 + 1 / ξ )⎥ (3-11)
⎣ 2 ⎦
in which ξ is the ratio of the distance d to the radius of the sphere
d
ξ= (3-12)
ra
If we take the radius ra to be 10 μm and the distance d to be 250 μm, then

the capacitance between the sphere and plate becomes 1.13·10-15 F.
The second analytical approximation is based on a model that is used in the

ink jet technology, where a liquid jet breaks up in droplets inside a hollow
cylindrical charging electrode. According to [36], the capacitance between the
droplet to be formed and the charging electrode is given by
69
2πε 0 l
C droplet = (3-13)
⎛D ⎞
ln⎜ c ⎟
⎜D ⎟
⎝ j⎠
in which Dc is the inner diameter of the charging electrode and Dj is the

diameter of the liquid jet, which relates to the droplet diameter as Dd = 1.89·Dj. The
value of l is set to the length of the jet forming one droplet, which in the case of
these particular HPMS devices is 4.3·Dj. With Dc = 500 μm and the droplet
diameter Dd = 20 μm, the capacitance between the charging electrode and the
droplet is 0.65·10-15 F.
Lastly, we can use the simulation software package LORENTZ to calculate

the capacitance between the droplet and the charging electrode as shown in Figure
3.14. This yielded a value of 0.83·10-15 F.
We can see that all three calculated capacitances are of the same order of
magnitude and that the values provided by the ink jet model and the simulation in
LORENTZ are close together. Since the situation as shown in Figure 3.14 is the
most realistic approximation of the spraying system, we will use the latter value for
the capacitance.
To calculate the potential of the charging electrode, we take the charge on a

droplet as calculated at the beginning of this section and use it in equation (3-10).
This yields that, theoretically, the charging electrode should be set on a potential of
229 V to obtain a water droplet of 20 μm with a charge of 1.9·10-13 C, which
should produce an output current of 0.9 μA per HPMS device as stated in the
beginning of this section.
3.4.3 HPMS implementation and experiments
It was already known that the HPMS system is capable of producing

monodisperse droplets and this has been confirmed by a high speed camera test set-
up, where uncharged droplets were sprayed, as shown in Figure 3.15.
70
Figure 3.15. Image captured with a high speed camera of the spraying with the HPMS
system. The spraying liquid is water+0.9% NaCl and the droplets are not yet charged.
The droplets in this image have a diameter of 20 μm.
To illustrate the effect of the charging process, the potential on the

charging electrode has been varied from 100 to 500 V. In Figure 3.16, where the
charging potential has been set to 300 V, we can see that the charged droplets now
repel each other. Another effect that was observed was that the repulsion between
the charged droplets became stronger and more visible as the charging potential
was increased.
Figure 3.16. The charging electrode is set to a potential of 300 V. The droplets are
sprayed from the left to the right. We can see that the droplets repel each other as
they move to the right. The droplets in this image have a diameter of 20 μm.
The current has been measured for a single HPMS device using a set-up as
shown in Figure 3.13. The potential of the charging electrode was set at 300 V and
71
this yielded a current of 0.1 μA. This is lower than the predicted current calculated
in the previous section by almost a factor of 10.
We can conclude that the droplets are not charged up to 30% of the
maximum charge as given by the Rayleigh limit, but only up to 3-4%. This implies
that the calculated capacitance is too high, which means that the electrode distance
is estimated too small.
Also, measurements have been performed with a negative potential,

creating positively charged droplets, on the charging electrode. For a single HPMS
device, no significant difference in current was measured. However, we will see in
chapter 4, where multiple HPMS devices are tested, and chapter 5, where these
HMPS devices are implemented in the EWICON system, that a negative polarity
results in slightly lower currents.
3.4.4 Conclusions on HPMS based spraying systems
From the images that have been taken with the high speed camera, we can
conclude that the sprayed droplets are indeed monodisperse. These images also
show that the droplets can be charged by induction, because the droplets start to
repel each other.
3.5 Suitable charging method for the EWICON

3.5.1 Summarising conclusions
Both spraying methods have been investigated and both have been able to
produce charged water droplets which are suitable for the EWICON system. While
the EHDA spraying method was not able to produce monodisperse droplets with
water as the spraying liquid, the associated current produced by electrospray was
still in the same order of magnitude as the current produced by the HPMS devices.
Testing with the EWICON will have to show whether the polydispersity of the
EHDA produced droplets will affect the transportability by the wind.
It must be noted, that at this point in time, the effective wind surface areas
for both EHDA and HPMS are in the same order, which allows that the currents
can be compared.
Furthermore, the droplet creation for EHDA process required very little
power, in the order of a few nanowatts per spraying nozzle. The power required for
the HPMS method was in the order of a few milliwatts, but this spraying method
did produce monodisperse droplets using saline water as the spraying liquid.
72
In addition, with both spraying methods, it was possible to control the

charge on a droplet to a certain degree. With the HPMS method, the charge could
be controlled independently from the droplet diameter. With the EHDA spraying
method, however, there is some interdependency between the droplet charge and
diameter.
3.5.2 Decision
Based on the conclusions in the last section, it was decided to implement

both spraying methods in the EWICON system for further testing.
It could very well be thinkable that there will be a variety of different

EWICON systems each incorporating a charging system that is best suited for its
location. An EWICON system that is intended for operation at sea might be
equipped with a HPMS based system, because of its favourable spraying
capabilities towards saline water. An EWICON system that is intended for land-
based operation, like e.g. on top of a building, might be equipped with an EHDA
based charging system. However, in both cases, water can be used and this means
that there is no risk to the environment.
73
Chapter 4 – EWICON system design
4
EWICON system design
“A common mistake that people make when trying to design something completely
foolproof is to underestimate the ingenuity of complete fools.”
– Douglas Adams
Now that we have taken a look at the different types of implementations of

the EWICON system in chapter 1 and the theoretical background of charged
droplets being generated and moved by the wind in an electric field in chapter 2,
we shift the attention to the design of an actual EWICON system. A number of
issues need to be addressed before a final EWICON system can be successful.
We will start in section 4.1 with the general design, which will be based on
the implementation Type B, described in chapter 1, combined with either EHDA or
HPMS as the spraying method, described in chapter 3. There will be differences in
the design depending on the used spraying method. This design will then be the
starting point on which further analysis will be performed. Also, it will be the blue
print of the actual EWICON system.
In section 4.2, we will test the EWICON design with EHDA and HPMS in
a wind environment in order to determine whether the wind is able to move the
charged droplets away. We will also investigate whether the polarity of the charged
droplets affects this transportation process.
The next issue is the increase of the produced current. In the ideal case, the
EWICON system is expected to convert all of the Betz-limited power in the wind.
Because the EWICON system acts as a current source, ideally, we want the
EWICON system to be able to deliver a current that is matched to the power in the
wind. This required current will have to be significantly higher than the current
produced by a single spraying nozzle system, regardless of the charging method.
Therefore, in the actual EWICON system, multiple spraying nozzles need to be
combined in order to increase the rate of production of charged droplets and, thus,
75
the current. In section 4.3 we will look at extensions of both the EHDA and HPMS
single nozzle systems.
Another issue, related to the movement of charged droplets, is the

minimum required wind speed. For a simple system as described in chapter 2, we
can determine analytically how much wind is needed to move the droplets against
the electric field. However, in a more complex setting, where the electric field can
not be calculated as straightforward, we have to resort to numerical methods. It
appears that by analysing the droplet trajectories, we can find ways to lower the
minimum required wind speed, thereby increasing the applicability of the
EWICON system. This will be discussed in section 4.4.
In section 4.5, we will discuss some remaining issues. In chapter 3, the

charged droplet creation methods have been analysed, assuming the charging
system to be connected to earth. In the actual EWICON system, the potential of the
whole system, including that of the charging system, will rise. We will describe
how this will affect the behaviour of the charged droplets compared to the
considerations made in chapter 2 and 3.
We will conclude with the final laboratory designs for the EWICON
system, which will be described in section 4.6. Using these designs, a number of
experimental set-ups were constructed, which were then used for experiments and
measurements. These will be discussed in the next chapter.
4.1 General design
In Figure 4.1, we again take a look at the schematic overview of a Type B

implementation, i.e. the EWICON system without a separate charge collector unit.
The first thing that is apparent is that we need a platform that is isolated from earth.
Basically, all the components on this platform will have their common earth point
at this platform. Using switches, the platform can either be connected directly to
earth or to an electrical load. In this chapter, we will look at the charge droplet
transportation while the EWICON system is still connected to earth, i.e. in the case
that the wind moves away the charged droplets, the platform and, thus, the
EWICON system will not be charged.
Also, not explicitly mentioned before, the liquid has to be fed to the
spraying nozzles, thus requiring some form of reservoir and a method of
transportation of the liquid.
Lastly, the electrical devices, required to supply the high voltage and the
liquid, need to be powered. Considering the fact that the charging system on the
76
platform is isolated from earth, a self-sufficient solution needs to be found for the
power supply.
Figure 4.1. Schematic overview of a Type B EWICON implementation. A reservoir

supplies the spraying liquid to the isolated charging system, where charged droplets
are produced. The wind carries the charged droplets away, thereby increasing the
potential of the EWICON system.
4.1.1 Platform and isolation
In both implementations of the EWICON system, during normal operation,

there will be an electrically floating part or component that will be charged. In the
Type A implementation, it is the separate collector unit and in the Type B
implementation, it is the isolated charging system. If the EWICON system is
operated without an electric load, the floating part that is charged can be
represented by a capacitor. In the case that there is an electric load, this can be
represented by a resistive impedance parallel to the capacitor.
Critical to the operation of a Type B implementation is the fact that the

charging system needs to be isolated from earth. In practice, this means that all the
additional components and devices have to be isolated from earth such that any
leakage currents to earth are minimised. Therefore, in our experimental test set-up,
a rectangular metal plate placed on a set of insulators was used as the basis for the
charging system. This plate was sufficiently large to hold all the necessary
components.
77
So, by using a metal plate as a holding platform, a capacitor was

introduced in to the EWICON system. The value of the capacitance of the platform
can be estimated by either using a parallel-plate approximation, by calculation in a
software package such as LORENTZ or by measurement. The expression for the
capacitance, in first approximation, is
A
C EWICON = ε 0 ⋅ (4-1)
d platform
in which A is the area of the platform and dplatform is the height above
ground. In this case, the surface area, A, is 2.45 metres by 2.00 metres and the
height, d, is 1.13 metres. Rounding off the result, we get an indicative capacitance
of 50 pF. However, equation (4-1) is only valid when d is small compared to the
dimensions of the plate and in this particular case, we can see that d is in the same
order as the surface area, A, of the plate. This means that, because the fringe fields
existing at the edges of the plates will contribute to the capacitance, the actual
value is likely to be higher.
Therefore, a platform with these dimensions was entered into simulation

package LORENTZ, see Figure 4.2. The capacitance calculated by LORENTZ was
approximately 125 pF. This was in agreement with experiments in which the AC
impedance or capacitive reactance, Xc, of the system was determined by applying
an AC voltage. The capacitance was then obtained by using
1
C EWICON = (4-2)
2π ⋅ f ⋅ X c
in which f is the frequency of the applied AC voltage, which in these

experiments was set at 50 Hz.
If the charging process is established and the wind moves all the droplets
away, the produced current will charge the EWICON, because the spraying nozzles
are electrically connected to the platform. We can use the estimated capacitance of
the system to calculate the rate at which the potential of the EWICON system will
increase when e.g. one spraying nozzle disperses charged droplets with an
equivalent current of 1 μA. This is, as we have seen in chapter 2 and 3, a typical
value for the current. The output potential is related to this charging current of 1
μA by the following equation
78
dV
I = C EWICON ⋅ (4-3)
dt
in which I is the charging current, CEWICON is the capacitance of the

EWICON system and V is the measured output voltage potential of the EWICON
system.
Figure 4.2. The capacitance of the platform, a metal plate (green), to the earth has
been calculated by LORENTZ. The grey area represents earth or, in this case, the
floor above which the platform floats.
From this equation, it follows that if the system is charged with 1 μA

without an electrical load for 10 seconds, the potential of the EWICON system will
reach 80 kV. Therefore, porcelain support insulators were used, which exhibited
leakage currents in the several nanoamperes when stressed at DC potentials of 65
kV, implying a resistance in the order of tens of teraohms.
In the example, we assume that leakage currents present due to corona

discharges or due to currents flowing over the surface of the support insulators are
negligible. Also, we assume that the spraying behaviour of the charging system, i.e.
the nozzle/electrodes configuration, remains unaffected and that the wind is still
able to move the droplets away from the system.
79
4.1.2 Liquid supply system
In the previous chapter, we discussed the two main charging methods and
found that typical flow rates vary between 1 and 20 ml/hr per spraying nozzle or
pore. If we assume the output power per nozzle to be 50 mW, as we theoretically
deduced in chapter 2, then for a 1 W rated system we would require a flow rate of
400 ml/hr. For a 1 kW rated system, a flow rate of 400 litres per hour would be
required.
In the smaller systems, these flow rates can be provided by syringe pumps.
However, as soon as we enter the kilowatts region, the liquid supply would have to
be realised by pumping the liquid to a reservoir at a certain height, depending on
the placement of the EWICON system. From the reservoir, the liquid flows to the
spraying nozzles under the influence of hydrostatic pressure. The power, Ppump,
associated with pumping liquid to a certain height, h, is given by
dm
Ppump = ⋅h⋅ g (4-4)
dt
in which dm/dt is the mass flow rate of the liquid and g is the gravity
acceleration. As an example, we use water as the spraying liquid and this water
needs to be pumped to a height, h, of 10 metres at a flow rate of 400 litres per hour.
In that case, the minimum required pumping power will be 11 W, which is roughly
one percent of the rated power. We also have to keep in mind that equation (4-4)
only holds when the pump has a (theoretical) efficiency of a 100%. Currently, the
maximum efficiency of water pumps is roughly 90% which means that the required
power will actually be 12 W.
However, if the charging and spraying efficiency is lower and the output
power per nozzle is, for example, 5 mW or 0.5 mW at the same flow rate, then the
minimum required pumping power would be 121 W or 1.2 kW respectively. In the
latter case, it is clear that the EWICON will not be able to convert wind power to
electrical power with a positive efficiency. Therefore, in order for the EWICON to
be self-sustaining, the chosen charging process needs to be optimised to the extent
that the power associated with the required amount of spraying liquid is lower than
the output power of the EWICON system. Another possibility is to use liquid that
has reached the required height due to other processes, e.g. precipitation. We will
touch upon this subject in chapter 6 on recommendations.
In our experimental set-up, however, we have chosen to use a syringe

pump that can supply liquid to spraying nozzles in either an EHDA or an HPMS
80
based charging system. In this way, for research purposes, the liquid supply can be
accurately established to the desired flow rate.
4.1.3 Electrical system
As said before, the Type B implementation of the EWICON system

requires the charging system to be isolated from earth. This means that
conventional means of electricity supply to the system cannot be used, since this
would require an electric connection between the EWICON system and earth. This
connection, in turn, would mean that the accumulated charge would flow back to
earth and the EWICON system would be unable to drive a load.
Therefore, in order for the equipment in the EWICON system to operate,

power is required, which has to be generated on the EWICON platform itself. In
our experimental test set-up, we use a 12 V battery placed on the platform. Using a
DC/AC converter, the voltage is converted to 230 VAC, which is required by all
the components on the EWICON platform.
Among these components are the high voltage DC sources powering the
electrodes, which are responsible for the electric field required to establish the
spraying and charging process. These DC sources operate at potentials varying
from ±2.0 to ±14.0 kV and are connected to the platform, just like the spraying
nozzles. The reference potential of the platform is floating. Therefore, when the
potential of the EWICON system rises, the potential of the electrodes and the
spraying nozzles are lifted equally and, thus, the electric field in the vicinity of
spraying nozzle remains constant.
4.2 Charged droplet transportation
The isolated platform, the liquid pump and the independent electrical
system discussed in the previous section provide the basis on which the charging
system will be placed. We now proceed to determine whether the wind is able to
move the charged droplets away from the charging system. The current associated
with the droplets that are moved away will be called the displaced current. Firstly,
the wind generator will be described, after which the results of the implementations
of both spraying methods, discussed in chapter 3, will be discussed.
In chapter 3, the various implementations of both charging methods were

analysed in terms of produced currents associated with the charged droplets and, in
the ideal case, all of these charged droplets will be moved away by the wind. If this
is the case, then no charged droplets will come in contact with any of the electrodes
81
or the metal platform and, thus, no currents will be measured flowing back from
the electrodes to the high voltage power supply.
If any of the charged droplets are attracted back to e.g. the charging
electrode, then a flow-back current will be measured. Comparing these flow-back
currents with the produced currents will then give an indication of the effectiveness
of the wind to overcome the electric field.
4.2.1 Wind generator
The air flow in this research has been provided by a wind generator, as
shown in Figure 4.3, which basically consists of a fan that forces the air through a
system of small tubes. In this way, a laminar flow of air is obtained, which means
that the wind speed at each of the spraying nozzles is roughly equal.
The wind speeds could be varied from 2 to 14 m/s (2 to 7 on the scale of

Beaufort) for a circular wind surface area with a diameter of 35 cm. During the
testing of the implementations discussed in chapter 3, the wind speed was first set
at 10 m/s and then increased to 12 m/s.
In the early stages of the research project, a smaller and less powerful wind
turbine was available with a maximum wind speed of 8 m/s. Therefore, the
experiments concerning the Delrin nozzle in this chapter and chapter 5 have been
conducted with 6 and 8 m/s.
Figure 4.3. The wind generator which has been used to provide laminar air flow to
move the charged droplets away from the charging system.
82
4.2.2 EHDA
• Delrin nozzle
The first implementation tested was the Delrin nozzle

using a 70%/30% water/ethanol mixture which was configured to
yield a produced current, IN, of 0.5 μA. The measuring set-up has
been depicted schematically in Figure 4.4.
Figure 4.4. A schematic overview of the measuring test set-up for electrospraying
charged droplets. The wind is blowing from the left to the right. Liquid is fed to the
spraying nozzle using a pump. A high voltage DC source is used to power the charging
electrode, CE. The current associated with the production of charged droplets, IN, is
measured using an ammeter. Any charged droplets that will be attracted back to the
charging electrode, CE, will be detected as a flow-back current, ICE.
After the air flow was established, a flow-back current of 0.2 μA was
measured. This implies that the wind was able to displace a current of 0.3 μA
associated with the charged droplets. After the increase of the wind speed, the
flow-back current decreased to 0.1 μA.
When pure demineralised water was used as the spraying liquid, the
produced current was approximately 0.1 μA. After the air flow was established, a
flow-back current of 0.1 μA was measured, implying that nearly all of the droplets
were attracted back to the charging system. This, in turn, would mean that wind
83
was practically not able to displace a current. In this case, an increase of the wind
speed did not result in a decrease of the flow-back current.
Part of the droplets created by the electrospraying of water are highly

charged submicron droplets and even with the increased wind speed, the electric
force pulling the droplets back is higher than the wind drag force.
• Cylindrical electrode/nozzle configuration
The second implementation that was tested was the

cylindrical electrode/nozzle configuration, which was specifically
designed to spray demineralised water. The measured current
produced by the spraying process was found to be 0.3 μA. After the
air flow was established, a flow-back current of 0.1 μA was
measured. This implies that the wind was able to displace a current of
0.2 μA associated with the charged droplets. After the increase of the
wind speed, the flow-back current did not decrease.
Even though the electrospraying of water was improved with this new
configuration, there are probably still a number of submicron droplets present.
Similar to the situation described in the previous section, they cannot be moved by
the wind because the electric force is higher due to the relative high charge on the
small droplets.
• Polarity of the charged droplets: EHDA
As we have found in chapter 3, the polarity of the charging electrodes did

not affect the magnitude of the produced current. During these wind experiments,
the charging electrodes were set to a negative potential and this means that the
droplets are positively charged. However, when the charging electrode was set to a
positive potential, meaning the droplets were negatively charged, the flow-back
current was found to be equal to the produced current. This implies that the wind
was unable to move any of the charged droplets away from the EWICON system.
This also implies that the EWICON system will not be charged, when the
connection to earth will be cut. We will verify this in chapter 5.
This effect was observed for all implementations that used the EHDA
method, which also include the (still to be discussed) scaled-up versions. Also, this
effect was observed with all tested spraying liquids, i.e. water, water/ethanol
mixtures and ethanol.
According to Figure 3.5, a positive potential results in the presence of a

fine mist, while this mist is absent in the case of a negative potential. While this
84
effect is not yet fully understood, the reasons could lie in the fact that the creation
of the droplets is fundamentally different in both cases. When a positive potential is
used, the OH- ions are accelerated towards the apex of the cone. In the case of a
negative potential, it will be the H+ ions, which are smaller than OH- ions.
Therefore, it is likely that the droplets that break off of the cone due to H+ ions will
smaller in diameter.
4.2.3 HPMS1
• Single HPMS device
In chapter 3, a single HPMS device yielded a current of

0.09 μA associated with the charged droplets. After the air
flow was established, it was observed that all the droplets
were moved away from the EWICON system by the wind. This implies that all the
charged droplets can contribute to the charging of the EWICON system.
• Polarity of the charged droplets: HPMS
During the wind experiments with the HPMS method, no flow-back

current was measured regardless of the polarity of the droplets.
4.2.4 Conclusions: charged droplet transportation of both

methods
For the single nozzle implementations of both spraying methods, we can

conclude that the produced charged droplets can be moved by the wind against the
electric field generated by the EWICON system. In the case of the EHDA method,
the effectiveness depends on the sprayed liquid, the implementations and the
polarity of the droplets, unlike in the case of the HPMS method.
With respect to the EHDA method, in terms of displaced currents, the

increase of the wind speed in these single nozzle systems results in a higher current
when water/ethanol mixtures are used as the spraying liquid. If demineralised water
is used, the increase of the wind speed does not result in an increased displaced
current, because of the submicron sized droplets.
1
85
4.3 Increased charged droplet production
The projected output power of a single spraying nozzle is in the order of

tens of milliwatts, depending on which spraying method is used. Therefore, in
order to increase the output power of the charging system to the order of kilowatts,
the output current and, thus, the rate of charged droplet production needs to be
significantly increased.
One straightforward method to increase the production of droplets is to

simply connect multiple nozzles together. We will take a look at one- and two-
dimensional expansions of the needle nozzle spraying system for EHDA. With this
increased current production, there will also be more flow-back current. We will
also investigate the net current as a function of the wind speed.
Subsequently, we will look at a spraying method that is based on the

EHDA principle, but has no need for individual spraying needle nozzles. Finally,
we will discuss the expansion of an HPMS system and the results of combining
several of the spraying chips.
The experiments that will be discussed in this section are to determine the
suitability of different concepts for up scaling droplet production in the EWICON
system:
• EHDA: Multi-needle systems
• EHDA: Self-adjusting wire spraying system
• HPMS: Multi-chip system
4.3.1 EHDA: Multi-needle systems
The single nozzle version of the EHDA system was of the Delrin type and
in order to obtain a multi-nozzle version of this system, a number of these casings
would have to be placed next to each other. However, due to the size of each
casing, a diameter of 3 cm, there would be room for only a limited number of
Delrin nozzles in a small space.
Another strategy to designing a multi-nozzle system is to take only the

spraying needles and place these in a grid. This could be achieved in numerous
ways and in Figure 4.5 we have depicted one possible configuration in which the
number of nozzles was expanded in one dimension.
86
Figure 4.5. Schematic overview of a possible multi-nozzle set-up configuration. The

spraying is fed through tubes to the needles (1) where EHDA spraying is achieved
through the use of rod electrodes (2). (a) side view (b) front view.
Here, a row of nozzles have been positioned in an insulating material. The

distance between the nozzles could be varied from 1.0 to 4.0 cm. Instead of
individual ring electrodes for each needle, the electric field needed to establish the
spraying process is now generated by two rod electrodes placed on both sides of
the nozzles as depicted in Figure 4.5. A nine needle multi-nozzle configuration can
be seen in Figure 4.6 where the spraying needles are 2.0 cm apart.
Figure 4.6. A nine needle multi-nozzle configuration using rod electrodes as the
charging electrodes. In this configuration, the spraying needles are 2.0 cm apart.
Calculation of the electric field underneath the needles showed that without
electrospraying the field increased with higher separation distances. This suggests
that the current would be higher with a separation of 4.0 cm compared to 1.0 cm.
The experiments with the configuration shown in Figure 4.6 have been
conducted using the same flow rate per nozzle and charging potentials as the single
nozzle configuration. The flow rate was 20 ml/hr and the charging potential was -
4.0 kV. Also, the separation distance has been varied from 1.0 to 4.0 cm.
87
The main conclusion is that the increase of the number of spraying nozzles
led to an increase in the produced current as compared to a single nozzle in the
same system. However, the current produced by each single nozzle was
significantly lower compared to the single nozzle/ring electrode configuration. The
maximum current per nozzle was 0.2 μA compared to 0.5 μA measured with the
single needle nozzle configuration in the previous chapter. These results were
obtained with separation distances of 2.0 and 4.0 cm.
When the separation distance was decreased to 1.0 cm, the current was
measured to be lower per nozzle, 0.1 μA. This is due to fact that the space charge
of the sprayed droplets negatively affects the spraying process of the neighbouring
nozzles. Therefore, a separation of 2.0 cm was chosen for subsequent multi-nozzle
experiments.
• Currents and electric field analysis
One part of the explanation for this behaviour is the fact that the electric
field generated by the rod electrodes is not as strong on the tip on the spraying
needle nozzle as compared to the electric field generated by the ring electrode.
Both systems have been modelled in LORENTZ with the electrodes set at -4.0 kV,
after which the electric field right underneath the needle tip was found to be 2.6
kV/mm in the ring electrode set-up and 2.2 kV/mm in the rod electrode set-up. In
these simulations the space charge effects of the droplets have not been included.
This problem can be partly resolved by increasing the potential on the rod
electrodes to e.g. -6.0 kV, which increases the electric field to 3.3 kV/mm. This
increased the current per nozzle to 0.3 μA while occasionally reaching 0.4 μA,
which still is lower than the current yielded by the single nozzle configuration.
Another part of the explanation could lie in the fact that the electric field
underneath the needle tip is not rotationally symmetric compared to the ring
electrode configuration. This is illustrated in Figure 4.7 and Figure 4.8.
Figure 4.7. Both the ring electrode configuration (left) and the multi-nozzle with rod
electrodes configuration (right) have been modelled in LORENTZ. The small orange
coloured square indicate in which plane the electric field have been plotted.
88
In Figure 4.7 we can see the overall situation of both implementations,

where the small orange coloured squares indicate in which plane the electric field
has been plotted. These plots have been enlarged in Figure 4.8, in which Figure
4.8a is a plot of the field underneath the needle tip in the ring electrode
configuration in Y-Z plane. Figure 4.8b is a plot in the same plane of the field
underneath the needle tip in the rod electrode configuration. Both electric field
plots are comparable in shape and magnitude.
Figure 4.8. Plots of the electric field underneath the needle tip (a) in the ring electrode
configuration in the Y-Z plane, (b) in the multi-nozzle rod electrodes configuration
both in the Y-Z plane and (c) in the X-Z plane (right).
However, if we look at Figure 4.8c, which shows the electric field in the X-
Z plane parallel to the rod electrodes, we can clearly see that the magnitude differ
from the other two plots. Overall, this means that the electric forces acting on the
ions in the Taylor cone are not symmetrical compared to the situation with the ring
electrode and this could mean that the break-off process of the charged droplets is
negatively affected. In [37] on the disintegration of water droplets, it is discussed
that, theoretically, the formation of a Taylor cone can only occur when the electric
field is azimuthally symmetrical. While the conditions that enable the forming of a
Taylor cone are not a prerequisite for electrospraying in general, the fact that these
conditions are not present with rod electrodes means that a lower current is to be
expected.
Experiments with ring and cylindrical electrodes have proven to be more

effective for EHDA based spraying and they seem to at least give credence to the
above-mentioned discussion, see [38] and [39].
89
• Non-linearity and two-dimensional system
A concerning observation was the fact that while the produced currents
could be added together, this process was not linear. The nine nozzle configuration,
depicted in Figure 4.6, did not yield 2.7 μA as one would expect based on the
individual nozzle result, but slowly reached 1.5 μA as the number of nozzles was
increased.
In order to further investigate this non-linear effect, a two-dimensional

charging and spraying system was constructed as shown in Figure 4.9. This system
comprises of 3 rows similar to the one-dimensional multi-nozzle system each
equipped with 8 spraying needles totalling 24 spraying units. Each nozzle had been
individually tested to yield 0.3 μA. In theory, this would mean that a total output
current of 7.2 μA could be produced with this spraying system.
Figure 4.9. An EHDA based charging test set-up equipped with a two-dimensional
configuration of spraying nozzles. This spraying set-up employed 24 needles.
However, non-linearity was also present in this system. Each of the rows
yielded a current of 1.5-1.6 μA and any combination of two rows yielded currents
between 2.3 and 2.7 μA. Finally, the full system yielded a current of 3.5 μA, which
is roughly half of the expected output current.
90
• Multi-needle with rounded ring electrodes system
A multi-needle system has been constructed in which each of the needles

has its own ring electrode similar to the Delrin nozzle configuration, as depicted in
Figure 4.10. Five rings have been shown, but the actual implementation was
designed to fit up to 16 needles. These ring electrodes have been rounded in order
to prevent corona discharges. This has been done to investigate whether the non-
linearity also occurs if the electric field around each spraying needle is
symmetrical.
Figure 4.10. Schematic overview of the multi-needle with ring electrodes system. Each
ring has an inner diameter of 20 millimetres. Five rings have been shown, but the
actual implementation was designed to fit up to 16 needles. (a) bottom view (b) front
view, the wind comes out of the paper.
This set-up has been tested using parameters similar to the previous
systems. The sprayed liquid was a 70%/30% water/ethanol mixture with flow rate
of 20 ml/hr. The wind speed was 10 m/s. One ring/needle configuration yielded a
produced current of 0.5 μA. When 4 needles were connected, the produced current
was found to be 2.0 μA. Thus far, the produced current is proportional to the
number of ring/needle configuration. Connecting a 5th needle resulted in a total
produced current of 2.3 μA, which is still within measuring inaccuracies. However,
connecting a 6th needle resulted in a current of 2.5 μA, thus bring the current per
needle to 0.4 μA. We will discuss these effects at the end of this chapter.
Up until this system, all the discussed multi-nozzle charging systems were
tested with water/ethanol mixtures. They have been tested with demineralised
water, but with little success in terms of produced current and stability of the spray
91
itself. However, spraying demineralised water in a 6 nozzle version of this system

yielded a net current of 0.5 μA.
• Multiple needle with cylindrical electrode configurations
A three nozzle version of the cylindrical electrode for

spraying demineralised water, described in chapter 3, has also
been constructed. In section 4.2.2, we found that for one nozzle,
the produced current was 0.3 μA and the flow-back current was
0.1 μA. Therefore, we expect that the produced current would be 0.9 μA and that
the flow-back current would be 0.3 μA, yielding an expected displaced current of
0.6 μA.
During the measurements, the flow-back current was found to be 0.6 μA,
reducing the displaced current to 0.3 μA. Apparently, non-linearity is also present
in this case, however, this time, the non-linearity occurs due to the higher flow-
back current. This could be attributed to the fact that charged droplets are pushed
back by other charged droplets that have been sprayed a moment earlier. The fact,
that there are 3 instead of 1 spraying nozzles, means that the charged droplets
sprayed by the middle nozzle have less room to disperse.
This implies that in order to decrease the flow-back current, the wind speed
should be increased. Therefore, the wind speed was increased from 10 to 12 m/s.
This led to the flow-back current decreasing minimally by 0.1 μA to 0.5 μA and an
increase of the displaced current increased to 0.4 μA. However, if we allow for
measurement inaccuracies, then the expected and actual displaced current do not
differ by the same measure as the previously discussed multi-nozzle systems.
• The displaced current vs. wind speed
In the previous experiments with the single nozzle systems, we have

already investigated whether an increase in the wind speed of 2 m/s would reduce
the flow-back current. This has also been investigated for one of the multi-nozzle
systems. In this case, a four needle version with the rounded ring electrodes
configuration was tested.
The wind speed was varied from 6 to 12 m/s and the potential of charging
electrode was increased such that the flow-back current was increased to be 1.4 μA
when the produced current at the nozzles was 1.7 μA. The results can be seen in
Figure 4.11. As the wind speed increased, the flow-back current steadily decreases
to 0.4 μA. From this we can conclude, that the produced droplets vary in size.
92
displaced current vs. wind speed
2,0
1,8
1,6
1,4
net current (µA)
1,2
displaced current (µA)
1,0 nozzles current (µA)
0,8 flow-back current (µA)
0,6
0,4
0,2
0,0
0,0 2,0 4,0 6,0 8,0 10,0 12,0 14,0
wind speed (m/s)
Figure 4.11. The displaced current as a function of the wind speed. The current
measured at the nozzles stays roughly constant while the flow-back current decreases.
Thus, the displaced current increases with increasing wind speed.
Another effect that can be observed in Figure 4.11 is the fact that the
nozzles current increases at 10 m/s. This is due to the fact that at this wind speed,
sufficient charged droplets are moved away from the nozzles and this leads to an
increase in the local electric field. This, in turn, leads to the creation of droplets
with a higher charge and thus a higher produced current.
4.3.2 EHDA: Self-adjusting nozzles wire spraying system
As we have seen, the expansion of the needle nozzles system leads to a

higher, albeit at this time not proportional, current. The drawback is that this also
means that more spraying needles and electrodes need to be implemented in the
system. Concurrent with the research and design of multi-nozzle systems, research
has also been conducted into other ways of spraying charged droplets. One of these
methods, called high voltage wire spraying, is based on the observed phenomenon,
that when a wet wire is set on a high potential, small EHDA-based Taylor cones
start to appear.
For example, after rainfall, the wetted high voltage overhead lines appear
to spray water droplets. This is schematically depicted in Figure 4.12.
93
Figure 4.12.Taylor cones formed as on a high voltage overhead line.
The main advantage of high voltage wire spraying is that the Taylor cones
have been observed to be self-adjusting, i.e. because of the interaction between the
liquid and the electric field generated by the electrodes and the charged droplets,
the number of and the distance between the Taylor cones is adjusted according to
parameters such as flow rate, electric field and wind speed. This ensures the
optimal number and spacing of spraying nozzles which is not easily achieved using
needles.
There are several ways to implement this wire-spraying technique, but we

will discuss the final implementation which has been tested with ethanol as the
spraying liquid, see Figure 4.13. A spraying system was developed which sprays
charged droplets from Taylor cones formed from a liquid surface. This liquid
surface is created by a strip of porous material in which an electrode is integrated,
the length of this strip is 160 mm and the width is 4.5 mm. A number of standard
spraying needles (not shown in Figure 4.13 but they are in Figure 4.14) are
providing the liquid for the porous strip.
These porous strips were cut out of fluidising sheets manufactured by the
company Porex.
94
Figure 4.13. Schematic overview, taken from [40], of the self-adjusting multi-nozzle
set-up, with and without applied voltage. Liquid is supplied to the porous strip, which
evenly distributes it to a liquid film. Under the influence of gravity suspended droplets
appear, see upper part. After the voltage has been applied, Taylor cones appear, see
lower part.
Experiments have been conducted with this spraying system to test whether
charged droplets would be created and whether the self-adjusting properties would
be observed. In Figure 4.14, we can see a photo of the spraying process.
95
Figure 4.14. Photos of experiments with the self-adjusting multi-nozzle set-up. On the
left-hand side, we can see the needle nozzles supplying liquid to the white porous strip
from which the Taylor cones are created. On the right-hand side, we can see an
enlargement of one these Taylor cones.
Using this system, the flow rate was set at 120 ml/hr, which is comparable
to the flow rates in EWICON systems using multi-nozzle EHDA spraying. The
charging potential was varied from 9.0 to 15.0 kV, which resulted in a charged
droplet current increasing from 1.5 to 6.2 μA. The number of self-adjusting
spraying sites went from 6 to 13. Another observed effect, while increasing the
charging potential, was that the size, i.e. the length and width, of the Taylor cones
decreased when their number increased, as depicted in Figure 4.15.
Figure 4.15. Increasing the charging potential leads to an increase of number of

Taylor cones. Since the flow rate remains constant, the dimensions of the cones
become smaller as schematically depicted.
Following these tests, the wind generator was turned on at a speed of 12

m/s. The first effect that could be observed was that the Taylor cones were strongly
deformed by the wind. The second effect that could be observed was that the
distance between the Taylor cones decreased. This could be explained by the fact
that the wind transports charged droplets away, thereby effectively removing space
charge from the vicinity of the cones. Normally, the presence of the charged
droplets underneath would be responsible for the repulsion between the cones.
However, with the removal of the droplets by the wind, the repulsive force
decreases and, thus, the distance between the cones.
96
Theoretically, this would leave room for more cones, which could be
achieved by increasing the flow rate at this point, thus increasing the produced
current. However, due to safety reasons, this has not been tested and the spraying
system as a whole needs to be modified such that the flow rate can safely be
increased during the spraying process.
We can thus conclude that the Taylor cones are self-adjusting and will
adapt to the prevailing parameters such as wind speed and charging potential.
4.3.3 HPMS: Multi-device systems1
Three of the HPMS devices, each containing 35 spraying pores, discussed

in the previous chapter have been combined in a way similar to the EHDA-based
spraying needle nozzles. The output current of this three HPMS device spraying
system has been measured to be 0.3 μA as compared to 0.09 μA for the single
HPMS device spraying system. Therefore, the conclusion can be drawn that with
respect to the number of spraying pores, the output current scales linearly.
4.4 Lowering required wind speed by field grading
In this section, we will investigate whether it is possible to grade the

electric field in such a way, that the minimum required wind speed for effective
EWICON operation can be lowered, while still maintaining the charged droplet
creation process. To this end, the electric field is analysed and the droplet
trajectories are calculated at different wind speeds. Subsequently, the electric field
will be modified using one or more extra electrodes, after which the droplet
trajectories will be plotted again.
As an example, we will look at a charging system that consists of the ring

electrode with needle nozzle set-up used for the EHDA spraying. For HPMS at this
moment, there is no need to lower the minimum required wind speed, because of
the high exit velocities of the charged droplets.
4.4.1 Analysis of droplet trajectories
In Figure 4.16, the trajectories of droplets have been plotted as function of

the wind speed, which is directed in the positive Y-direction and varied from 6 to
10 m/s. The needle nozzle is set at 0.0 kV, representing the fact that EWICON
system is connected to earth, and the ring electrode is set at -4.0 kV. As we can see,
the droplet escapes the charging system only when the wind speed is 9 m/s or
1
97
higher. Also, we can see that at 8 m/s, the wind manages to move the droplet, but
still has insufficient time to accelerate the droplet to its terminal escape velocity.
Figure 4.16. The trajectories of droplets are plotted as a function of the wind speed.
The ring electrode is set at -4.0 kV and the needle is set at 0 kV. The wind speeds in
this plot vary from 6 to 10 m/s. The droplet is moved away completely, only when the
wind speed is 9 m/s or higher.
It should be noted, that in these simulations with LORENTZ, the wind is

modelled as laminar air flow, similar to the MATLAB modelling performed in
chapter 2.
Another example, which we will explore more in detail, is a charging

system that uses the multi-nozzle spraying system described in section 4.3.1. In
Figure 4.17, a five nozzle system with two metal rods as the charging electrodes
has been depicted. Another component that is visible in the right lower corner of
Figure 4.17 is a metal plate representing the platform as described in section 4.1.1.
The orientation and placement of the platform depends on the actual
implementation of a particular EWICON system.
The wind speed is set at 6 m/s and pointed in the positive Y-direction. The
two rod electrodes are set at -4.0 kV and the nozzles and the platform are set at 0.0
kV, since they are electrically connected. For this simulation, trajectories of
multiple droplets have been plotted.
98
Figure 4.17. Result of the simulation of the EWICON system using a wind speed of 6
m/s. The two rod electrodes are set at -4.0 kV. The rectangular electrode represents
the platform of the grounded EWICON system and is electrically connected to the
needles. Most of the droplets are forced back to the charging electrodes.
As we can see in Figure 4.17, using a wind speed of 6 m/s (3 on the scale
of Beaufort), the wind was unable to move all of the droplets away from the
system. This has been verified in experiments where the leakage currents have
been measured at the charging electrodes.
Revisiting the calculations in chapter 2, we found that if the wind speed

was 3.3 m/s or more, then a charged droplet would be moved away by the wind,
provided the uniform electric field nowhere is higher than 104 V/m. However, in
this case, the volume underneath the spraying needles contains regions that have
electric fields varying between 104 and 105 V/m. Also, the fact that there is a
multitude of charged droplets present instead of one should be taken into account.
Together, these are contributing factors to the fact that charged droplets are
attracted back, even though the wind was assumed to be sufficient with a speed of
6 m/s.
Only when the wind speed was increased to 12 m/s (6 on the scale of
Beaufort), which is even more than was needed in the case of the single needle
nozzle/ring electrode configuration, the wind was able to move all the droplets
away from the system, as is shown in Figure 4.18.
99
m/s. Only the presence of a much higher wind speed than 6 m/s results in the complete
removal of all charged droplets.
In order to analyse in what way the electric field acts on the charged
droplets, we take a look at a cross section in the Z-Y plane in an area around the
spraying nozzles. We plot the z-component of the electric field, which results in
Figure 4.19.
Figure 4.19. Plot of the z-component of the electric field present in the EWICON
system. In the yellow/orange areas, the charged droplets are pulled downwards. In the
dark blue areas, the charged droplets are pulled upwards.
100
In this plot, the yellow/orange areas represent the parts where the electric
field is pointed downwards and, thus, positively charged droplets will be pulled
downwards, which is consistent with the mechanism of EHDA droplet creation.
The blue areas represent the parts where the electric field is pointed upwards and in
these regions, the droplets will be pulled upwards one of the two rod electrodes.
This is the effect we want to minimise. The underlying thought is to grade the
electric field in such a way, that the wind actually has more time to act on the
droplets before they can reach one of the electrodes.
4.4.2 Steering electrode(s)
One possible way of implementing that strategy is by introducing one or

more steering electrodes, on a potential with a polarity similar to the charging
electrode. The main thought is this steering electrode will change the electric field
in such a way, that the electric force on the charged droplets pointing towards the
charging electrodes will be lower. This means that the charged droplets will be
attracted towards the steering electrode rather than the charging electrodes,
travelling a different path. Also, this implies that in the case that there is no wind,
most of the charged droplets would be pulled to the steering electrode.
In Figure 4.20, a plot of a simulation in LORENTZ has been made where

there is no wind present. Most of the charged droplets are now attracted to the
steering electrode, while a small remaining number of charged droplets are still
attracted to the charging electrode. This is due to the mutual repulsion of the
charged droplets, which causes some charged droplets to move in the +Y direction,
thus being farther away from the steering electrode.
Figure 4.20. Result of a simulation of a charging system with a steering electrode when
there is no wind present. Most of the charged droplets are attracted to the steering
electrode while a small number of droplets are still attracted to the charging electrode.
101
If the distance covered by this different path to the steering electrode is

longer than the path to the charging electrodes, then, presumably, the wind should
have more time to act on those charged droplets. This, in turn, increases the
likelihood that the charged droplets will escape the charging system and contribute
to the produced current.
In Figure 4.21, we have plotted the electric field around the charging
system. This time, a rectangular shaped steering electrode has been added to the
system in front of the charging system with respect to the direction of the wind. We
can see in the plot that by adding the steering electrode, the z-component of the
electric field has decreased in the area where the charged droplets exit the nozzles.
Therefore, the droplets will, first of all, travel a greater distance downwards and,
secondly, be pulled back in the negative y-direction because of the steering
electrode.
Figure 4.21. Plot of the z-component of the electric field present in the EWICON
system. A steering electrode has been added and set at a potential of -10.0 kV.
Because of the fact that the charged droplets have to cover a greater
distance before possibly colliding with one of the charging electrodes, the wind has
more time to act on the droplets thus enabling it to remove most of the droplets
from the system. This has been simulated again, as can be seen in Figure 4.22, in
which the wind speed has again been set to 6 m/s.
102
m/s. In this system, a steering electrode has been added and set at -10.0 kV.
Practically, all of the charged droplets are now moved away by the wind.
We can see that practically all the droplets are removed from the charging
system, which again was in agreement with the experiments.
4.4.3 Conclusions on field grading
The main conclusion is that the addition of at least one steering electrode
already improves the usability of the EWICON system, with droplets created with
the EHDA spraying method. Therefore, in all the future implementations of the
EWICON, whether they are multi-needle or self-adjusting nozzle systems, steering
electrodes should be considered to lower the minimum required wind speed.
It must be noted that adding an extra steering electrode, in theory, does not
require any extra power after it has been charged up to its target potential, as long
as no charged droplets come in contact with that steering electrode. Adding extra
electrodes does, however, bring extra inconveniences with regard to the complexity
of the EWICON system. These electrodes operate at different potentials and this
means that the insulation between the various electrodes needs to be sufficiently
high as to prevent any leakage currents to occur.
Another observed effect was the fact that, during spraying experiments, the
potential needed on the charging electrodes was lower than if no steering electrode
was used. The electric field at the tip of the spraying nozzles is raised by the
electric field generated by the steering electrode. This is an advantage because this
decreases the possibility of discharges at the charging electrodes.
103
Overall, LORENTZ is a useful tool to aid in the design of an EWICON

system and to reasonably predict whether charged droplets will be moved away.
However, in these simulations, we assumed monodispersity and the experiments
have been carried out with water/ethanol mixtures, which exhibit monodispersity to
a certain degree. Therefore, experiments will still be needed to confirm whether
charged droplets will be moved away, especially when electrospraying of water is
employed, where monodispersity is not yet achieved.
4.5 Rising EWICON potential and droplet movement
During all the experiments discussed in this chapter, the EWICON system
itself was grounded while the charged droplets were sprayed and moved away by
the wind. However, as soon as the EWICON system is disconnected from earth, it
will charge up to a certain potential. As discussed in 4.1.3, the electric field in the
vicinity of the spraying system remains constant during the charging process,
because the potentials of the charging and steering electrodes will rise
correspondingly. Therefore, the spraying process, either based on EHDA or HPMS,
will remain undisturbed and the creation of charged droplets will continue.
This does not mean, however, that the electric forces acting on the droplets
will remain constant. As the potential of the EWICON system rises, the attractive
force back to the system is increased and at a certain point, the droplets will return
to system. This means that the flow-back current increases and that the charging
rate of the EWICON system decreases. Therefore, in the next chapter, in which we
will conduct the charging experiments, we expect to observe the charging current
to decrease while the charging potential will reach a maximum.
The behaviour still depends on the charging and spraying method, or more
specifically, the size and charge distribution of the droplets. In 3.1.3, we briefly
touched on the subject of monodispersity in that it implied that the force balance on
the monodisperse droplets would be the same. It also means that, theoretically, at a
certain potential, all the charged droplets would be attracted back to the system,
thereby instantly reducing the current to zero. We expect this for the HPMS
configuration.
Since EHDA does not provide monodisperse spraying for water and
water/ethanol, we expect the charged droplets with a higher electrical mobility to
be attracted back sooner than the charged droplets with a lower electrical mobility.
Therefore, the charging current is expected to decrease gradually as function of the
EWICON system potential.
104
4.6 Overall conclusions and final experimental set-up

4.6.1 Conclusions
We can conclude that it is possible to design the EWICON system in such

a way that the wind can move charged droplets away. In the case of EHDA based
spraying, the use of steering electrodes significantly lowers the minimum required
wind speed. In the case of HPMS based spraying, steering electrodes were not
needed, because of the high jetting velocity of the droplets.
Multiple-nozzle charging and spraying system, in one and two dimensions,

with both the EHDA and HPMS method were designed and constructed. It has
been possible to add the currents of the spraying nozzles together, even though for
EHDA based systems this adding process was not linear.
The reason(s) why this non-linearity is present in these systems are still a
topic of ongoing research and one of the most important issues that need to be
resolved in order for the EWICON system to be scalable. The currents have been
measured again more accurately using an oscilloscope instead of ammeters for a 1,
3 and 6 nozzles version of the rounded ring electrodes system.
The voltage has been measured over a 10 kΩ resistor resulting in Figure

4.23. We can see that in Figure 4.23a, that the voltage measured over the resistor
connected to one nozzle is 3.110 mV, whereas in Figure 4.23b, the voltage
measured over the resistor connected to three nozzles is 7.670 mV. Thus, the
current produced by one nozzle is 0.311 μA and the current produced by three
nozzles is 0.767 μA or 0.256 μA per nozzle.
Figure 4.23. Results of the voltage measured over a 10 kΩ resistor in a (a) single
nozzle (b) three nozzle (c) six nozzle version of the rounded ring electrodes system.
The three nozzle version only yielded 2.5 times the current produced by a single
nozzle version. The six nozzle version only yielded 4.1 times the current produced by
the single nozzle version.
105
In Figure 4.23c, with 6 nozzles, the voltage measured over the resistor
connected to 6 nozzles is 12.88 mV i.e. a nozzle current of 1.288 μA, which is
higher only by a factor of 4.1 compared to one nozzle. Therefore, by also using
more precise measurements, we can conclude that the increase of the current is not
proportional to the number of nozzles.
The reasons for this non-linearity may lie in the fundamental nature in
which the charge separation takes place. In the research of EHDA in general, the
focus has mainly been on spraying particles in a controlled and predictable manner
and less on the electrical implications that go hand in hand with the
electrospraying. Therefore, research should be conducted on the effect of the
different materials at the nozzle/liquid interface on the charge separation process.
In EHDA, the nozzle/liquid interface has always been a metal/water or
metal/ethanol interface. In HPMS, the interface was always a dielectric/saline
water interface and, up to now, no non-linearity has been found with HPMS.
However, this was beyond the scope of this research project.
4.6.2 Final experimental test set-up
As it has been stated before, there are many ways to design and implement
the EWICON system. For the majority of the conducted experiments described in
this and the next chapter, the test set-up that was used is shown in Figure 4.24.
In Figure 4.24, we can see the isolated platform on which the high voltage
DC sources have been placed. They are connected to the charging and the steering
electrodes via ammeters. Also, on the platform, a 12V battery and a 12/240
converter are present, which power the HVDC sources and the syringe pump.
In order to facilitate the changing of the different implementations of the

spraying systems, a separate isolated support structure has been constructed, seen
on the right-hand side of the figure. The wind generator has been placed closely
behind the support structure to ensure that the full width of the generator can be
used for air flow.
The copper spheres (1), have been placed to minimise losses due to corona
discharges, which would occur otherwise at higher EWICON system potentials.
These rounded shapes can be seen throughout the whole set-up and all serve the
same purpose. This rounding aspect is an issue that always needs to be considered
when a scaled up version of the EWICON system will be designed.
The EWICON system potential has been measured with an electrostatic

voltmeter (2). The potential is measured directly through electrostatic forces
instead by measuring the current. In this way, the current losses are minimised.
106
Figure 4.24. The final experimental test set-up of the EWICON system. On the left-
hand side, the metal plate acting as the platform can be seen. On the right-hand side,
the charging and spraying system can be seen.
107
Chapter 5 – Testing of the complete EWICON system
5
Testing of the complete EWICON system
“It doesn’t matter how beautiful your theory is, it doesn’t matter how smart you are. If
it doesn’t agree with experiment, it’s wrong.”
– Richard Feynman
In the previous chapters, the separate parts of the EWICON system, e.g. the
charging system, the liquid supply system and the holding platform and its
insulation have all been proven to operate as designed. In this chapter, experiments
will be discussed that have been conducted on complete EWICON systems. The
objective of these experiments was to determine whether the wind is able to move
the charged droplets away such that the droplets can reach earth and if, by
achieving that, the EWICON system can be charged.
Subsequently, upon having analysed the behaviour of the EWICON system

without a load, it will be determined whether the EWICON system can be used to
power an electrical load. We will analyse these results and determine whether the
various EWICON configurations exhibit the expected behaviour. Summarising
from the previous chapters, we expect the following:
• Using the EHDA method for charged droplet creation with water or
water/ethanol mixtures, the charging or displaced current should decrease
gradually as a function of the EWICON system potential.
• Using the HPMS method for charged droplet creation, the charging or
displaced current should decrease rapidly to zero at a certain EWICON
system potential.
• The maximum attained EWICON system potential should rise with

increasing wind speeds.
• The EWICON output power should rise with increasing wind speeds.
109
All the experiments have been performed using a Type B implementation

of the EWICON system, i.e. without a separate charge collector unit, see [41]. In
section 5.1, the experimental method will be discussed. We will explain in what
way the various types of currents and potentials were measured.
In the sections 5.2, 5.3 and 5.4, the experiments are described that were
performed using different spraying methods. Some of the results of earlier
measurements will be briefly stated, so that the results can be compared to results
of more recent experiments. In section 5.5, we will analyse the results of the
experiments and determine the efficiencies of the various configurations and in 5.6
we will summarise the conclusions.
In section 5.7, we will conclude this chapter with some considerations on

up scaling of the EWICON.
5.1 Experimental method
In order to determine whether the EWICON system could be charged, the

following steps were taken. Firstly, the wind turbine generator as described in
chapter 4 was switched on and set to the desired wind speed. Secondly, the
spraying and charging of the droplets was activated, in a similar way as described
in chapters 3 and 4, depending on the used spraying system.
During this spraying phase, the holding platform of the EWICON system
was still connected to earth. At this point, the optimal settings for the potentials on
the charging and possibly steering electrodes, which were found during the testing
described in chapter 4, were used. This also means that for different charging
systems, e.g. the single nozzle/ring electrode system or the multi-nozzle/rod
electrodes system, different potentials will be used. Therefore, it can be assumed
that, in the next sections, the specified potentials are the ones that produce the
highest net charging current.
In the next phase, the holding platform was disconnected from earth, thus
allowing the EWICON system to charge like a capacitor. During this charging
process, the EWICON system potential was measured as a function of time. Using
equation (4-2), we can indirectly estimate the magnitude of the charging current.
The same procedure was followed to test the EWICON system with an electrical
load attached. Finally, by using the results of these experiments, the output power
of the various EWICON systems was calculated.
110
5.2 EWICON with single nozzle EHDA spraying

systems
The first charging systems to be tested in the EWICON system are the
spraying set-ups that consist of one single needle nozzle with one charging
electrode. We will discuss the set-up equipped with the Delrin nozzle, the single
nozzle version of the needle/rounded ring configuration discussed in chapter 3 and
the set-up equipped with the cylindrical electrode, both of which were discussed in
chapter 2.
5.2.1 Single ring electrode
• Delrin nozzle (configuration #1 and #2)
The first experiment using the Delrin nozzle with the ring electrode
was performed using a 70%/30% water/ethanol mixture at a flow
rate of 20 ml/hr. No steering electrode was yet implemented in this
configuration. Similar to the Delrin nozzle experiments discussed
in chapter 4, the wind speed was set at 6 m/s and the ring electrode
was set on a potential of -7.0 kV, thus creating positively charged
water droplets. The measured current from the needle nozzle was
0.5 μA and the flow-back current to the ring electrode was 0.2 μA.
After disconnecting the holding platform from earth, the potential of the
EWICON system rose to a value of 6.0-6.5 kV in 5 seconds, where it remained
steady for as long as the wind flow was present and the liquid was supplied.
Connecting a load of 20.5 GΩ to the EWICON system resulted in a drop of
EWICON potential to 2.0 kV. The current measured through the load was 0.1 μA,
which means that the output power of this configuration resulted in 0.2 mW.
To illustrate the difference between positively and negatively charged

droplets with respect to the transportation of said droplets by the wind, the ring
electrode was set on a positive potential to create negatively charged droplets. The
other parameters were kept constant. This time, after disconnecting the holding
platform from earth, the potential of the system rose to approximately 1 kV in 90
seconds. Attaching the load to the EWICON system resulted in a drop of the
potential to zero. In other words, this means that the load of 20.5 GΩ is too low
compared to the internal impedance of the EWICON system. This effect has been
noticed and tested for all EHDA based configurations with water, water/ethanol
and pure ethanol as the spraying liquid. Therefore, from this point onwards, all the
discussed results will only concern the experiments in which positively charged
droplets were created and sprayed.
111
This experiment has also been performed with demineralised water. The
measured current from the needle nozzle was approximately 0.1 μA and the flow-
back current was also approximately 0.1 μA. After the platform was disconnected
from earth, the potential of the EWICON rose to a 1000 V in 120 seconds.
Connecting a 5.3, 10.3 or 20.5 GΩ load resulted in a drop of the potential to zero.
• Rounded ring electrode (configuration #3)
In this configuration, the holder with rounded ring electrodes has been
used with only one spraying needle. In this way, the result can be
compared to the results of configurations with multiple spraying
needles. The flow rate was set at 20 ml/hr and the spraying liquid was
a water/ethanol mixture of 70%/30%. The wind speed was set at 12 m/s. The
measured current from the needle was 0.4-0.5 μA.
EWICON system rose to 45.0 kV in 60 seconds. The charging current has been
indirectly determined using equation (4-3) and the progression of the EWICON
potential over time, as can be seen in Figure 5.1. The theoretical output power
reaches a maximum of 3.1 mW at 20 kV, which means that maximum power
transfer would occur if a 130 GΩ load would be used. This load would then be
equal to the internal source impedance of the current source.
rounded ring electrode, 1 nozzle, 70%/30% water/ethanol
0,400 3,50
0,350 3,00
0,300
2,50
charging current (µA)
output power (mW)
0,250
2,00
Current (µA)
0,200
Power (mW)
1,50
0,150
1,00
0,100
0,050 0,50
0,000 0,00
0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00 45,00 50,00
EWICON system potential (kV)
Figure 5.1. The indirectly measured charging currents and output power of a single
nozzle while using ring electrodes as charging electrodes.
The charging current, Ic, vs. the EWICON system potential, UEWICON, has
been fitted with an exponential function in MATLAB, see appendix D, yielding the
following expression for the charging current:
112
I c (U EWICON ) = 0.501 ⋅ 10 −6 exp(−62.31 ⋅ 10 −6 ⋅ U EWICON ) (5-1)
according to which Ic(0)=0.5 μA. This agrees with the current measured
from the single needle nozzle. Based on (5-1), the calculated curves for the
charging current and output power have been plotted, as shown in Figure 5.2.
MODEL: rounded ring electrode, 1 nozzle, 70%/30% water/ethanol
0,60 3,50
0,50 3,00
2,50
output power (mW)

0,40
2,00
Current (µA)
0,30
Power (mW)
1,50
0,20
1,00
0,10
0,50
0,00 0,00
0 5 10 15 20 25 30 35 40 45 50
Figure 5.2. The calculated curves of the charging current and output power for the
single needle nozzles version.
This suggests that the maximum power transfer will occur at 15 kV as

opposed to the 20 kV derived from Figure 5.1. This discrepancy is due to the
measurement error in the readout of the potential and the inaccuracy of the
estimated parameters of (5-1).
5.2.2 Single cylindrical electrode
• (configuration #4)
This spraying set-up was used together with a steering electrode. The
used liquid was demineralised water at a flow rate of 20 ml/hr. The
wind was set at 12 m/s. During this experiment, the cylindrical
charging electrode was set on a potential of -2.0 kV and the steering
electrode was set on -9.7 kV.
113
EWICON system rose to 35.0 kV in approximately 5 minutes. The charging
current has been indirectly determined as can be seen in Figure 5.3.
cylindrical electrode, 1 nozzle, demineralised water
0,040 0,40
0,035 0,35
0,030 0,30
output power (mW)

0,025 0,25
Current (µA)
0,020 0,20
Power (mW)
0,015 0,15
0,010 0,10
0,005 0,05
0,000 0,00
0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00
Figure 5.3. The indirectly measured charging current and output power of a single
nozzle/cylindrical electrode configuration. The spraying liquid is demineralised water.
5.3 EWICON with multiple nozzle EHDA spraying

systems
In this section, we will discuss the experiments performed with the

EWICON system employing various spraying systems in which multiple spraying
nozzles are present. In section 5.3.1, we will start with the multi-nozzle systems in
which the charging electrodes are either rod electrodes or rounded ring electrodes
as described in section 4.3.1.
Also described in section 4.3.1, is the multiple needle nozzle system with
the cylindrical electrode configuration, of which the experimental results will be
discussed in section 5.3.2. Finally, in section 5.3.3, the results of the final version
of the self-adjusting nozzles configuration will be discussed.
5.3.1 Multiple needles with rod or rounded ring electrodes
• Rod electrodes (configuration #5, #6 and #7)
Eight spraying nozzles were used and the flow rate was set
at 20 ml/hr per nozzle. The wind speed was set at 10 m/s.
The rod electrodes were set on -5.0 kV and the steering
electrode was set on -10.0 kV. Firstly, demineralised water
114
was used as the spraying liquid. This resulted in a net current just under 0.5 μA,
while the EWICON system was still connected to earth. After disconnecting the
holding platform, the EWICON system potential rose to 3.0 kV.
Secondly, the spraying liquid was changed to a 70%/30% water/ethanol

mixture. This yielded a net current of 1.2 μA, while the EWICON system was still
connected to earth. After disconnecting the holding platform from earth, the
EWICON system potential rose to 9.8 kV. Connecting this configuration to a load
of 20.5 GΩ yielded an output potential of 8.2 kV. The current measured through
the load was 0.4 μA, resulting in an output power of 3.3 mW or 0.4 mW per
nozzle.
Lastly, the two-dimensional spraying set-up described in 4.3.1 was used

with the 70%/30% water/ethanol mixture. This set-up employed 24 spraying
needles. This yielded a net current of 3.5 μA, while the EWICON system was still
connected to earth. Connecting the 20.5 GΩ load to this configuration yielded in an
output potential of 15.0 kV. The current measured through the load was 0.7 μA,
resulting in an output of 10.5 mW or 0.4 mW per nozzle.
• Rounded ring electrodes (configuration #8, #9, #10 and #11)
Six spraying nozzles were used and the flow rate was set at
20 ml/hr per nozzle. The spraying liquid was a 70%/30%
water/ethanol mixture. The wind speed was set at 12 m/s.
The ring electrodes were set on a potential of -4.0 kV and the steering electrode
was set on a potential of -10.0 kV. While the EWICON system still was connected
to earth, the current measured from the six needle nozzles was 1.5 μA.
After disconnecting the holding platform from earth, the EWICON system
potential rose to 45.0 kV in, on average, 30 seconds and ultimately reached 69.0
kV. Connecting this configuration to a load of 20.5 GΩ yielded an output potential
of 16.0 kV in 10 seconds. The measured current through the load was 0.8 μA and
this means that the output power of the EWICON system is 12.5 mW or 2.1 mW
per nozzle. Also, electrical loads of 10.3 and 5.3 GΩ were connected to this
configuration, which yielded an output potential of 9.7 kV with a current of 0.9 μA
and 5.7 kV with a current of 1.1 μA respectively.
The current has also been determined indirectly in Figure 5.4.
115
rounded ring electrodes, 6 nozzles, 70%/30% water/ethanol
0,600 12,00
0,500 10,00
output power (mW)

0,400 8,00
Current (µA)
0,300 6,00
Power (mW)
0,200 4,00
0,100 2,00
0,000 0,00
0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00 45,00 50,00
Figure 5.4. The indirectly measured charging currents and output power of six nozzles
while using ring electrodes as charging electrodes. The spraying liquid is a 70%/30%
water/ethanol mixture.
The charging current, Ic, vs. the EWICON system potential, UEWICON, has
been fitted with an exponential function in MATLAB, see appendix D, and this
yielded the following expression for the charging current:
I c (U EWICON ) = 1.581 ⋅ 10 −6 exp(−58.85 ⋅ 10 −6 ⋅ U EWICON ) (5-2)
according to which Ic(0)=1.6 μA. This agrees with the current measured
from the six needle nozzles. Based on (5-2), the calculated curves for the charging
current and output power have been plotted, as shown in Figure 5.5.
MODEL: rounded ring electrodes, 6 nozzles, 70%/30% water/ethanol
1,800 12,00
1,600
10,00
1,400
output power (mW)
1,200 8,00
1,000 Current (µA)

6,00
0,800 Power (mW)
0,600 4,00
0,400
2,00
0,200
0,000 0,00
0 5 10 15 20 25 30 35 40 45 50
Figure 5.5. The calculated curves of the charging current and output power for the six
needle nozzles version.
116
Similar to the single needle nozzle version, Figure 5.5 suggests that the
maximum power transfer of 9.8 mW will occur at 15 kV. This is not in
contradiction with the measurements, because in Figure 5.4, there is no maximum
of the output power visible. Thus, the maximum power transfer occurs at a lower
potential than 20 kV.
This experiment has also been performed with demineralised water. Six
needle nozzles spraying with a flow rate of 20 ml/hr per nozzle. While the
EWICON system still was connected to earth, the net current was 0.5 μA. After
disconnecting the holding platform from earth, the EWICON system potential rose
to 33 kV in 60 seconds. The charging process is shown in Figure 5.6.
rounded ring electrodes, 6 nozzles, demineralised water
0,300 3,50
0,250 3,00
2,50
output power (mW)

0,200
2,00
Current (µA)
0,150
Power (mW)
1,50
0,100
1,00
0,050 0,50
0,000 0,00
0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00
Figure 5.6. The indirectly measured charging currents and output power of six nozzles
while using ring electrodes as charging electrodes. The spraying liquid is
demineralised water.
As can be seen, the charging current and the output power that is produced
with demineralised water are lower as compared to the case when a water/ethanol
mixture is used.
5.3.2 Multiple needle with cylindrical electrode

configurations
• (configuration #12 and #13)
Three of these cylindrical electrode nozzles have been used with

a flow rate of 20 ml/hr per nozzle using demineralised water. The wind speed was
set at 12 m/s. Similar to the single version of this type of nozzle, the cylindrical
electrode was set on a potential of -2.0 kV and the steering electrode on -9.7 kV.
117
EWICON system rose to 35.0 kV in approximately 45 seconds, which is
significantly quicker than using a single cylindrical electrode nozzle. The charging
current has been determined as can be seen in Figure 5.7. The output power reaches
a maximum of 2.7 mW at 20 kV or 0.9 mW per nozzle.
cylindrical electrode, 3 nozzles, demineralised water
0,300 3,00
0,250 2,50
output power (mW)

0,200 2,00
Current (µA)
0,150 1,50
Power (mW)
0,100 1,00
0,050 0,50
0,000 0,00
0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00
Figure 5.7. The indirectly measured charging currents and output power of three
nozzle/cylindrical electrode configurations. The spraying liquid is demineralised
water.
cylindrical electrode, 3 nozzles, tap water
0,120 0,45
0,40
0,100
0,35
output power (mW)
0,080 0,30
0,25 Current (µA)

0,060
0,20 Power (mW)
0,040 0,15
0,10
0,020
0,05
0,000 0,00
0,00 2,00 4,00 6,00 8,00 10,00 12,00
Figure 5.8. The indirectly measured charging currents and output power of three
nozzle/cylindrical electrode configurations. The spraying liquid is tap water with a
conductivity of 456 μS/cm.
In order to illustrate the fact that the chaotic spraying which occurs when
tap water is used as the spraying liquid negatively affects the produced and
displaced current, the previous measurement has also been conducted while using
118
tap water as the spraying liquid. Tap water has a higher conductivity than
demineralised water, 456 μS/cm in this case, and this will cause chaotic and
unstable electrospraying. This will result in a lower charging current and, thus, in
lower EWICON system potentials, which can be observed in Figure 5.8.
5.3.3 Self-adjusting nozzles configuration
This configuration has only been tested with pure ethanol as the
spraying liquid at a flow rate of 120 ml/hr, which roughly
compares to six needle nozzles spraying at 20 ml/hr in the other EHDA
configurations. At a charging potential of -15.0 kV, the current associated with the
droplet creation process was 6.2 μA. No steering electrode was used in these
experiments.
However, as has been mentioned in the previous chapter, the thirteen

Taylor cones, which are responsible for the spraying of charged droplets, were
strongly deformed by the wind. This was especially noticeable when the holding
platform was disconnected from earth. The open output potential rose to 2.5 kV
and fluctuated heavily between 2 and 3 kV.
The self-adjusting nozzles configuration was not yet aerodynamically

optimised for the EWICON system. If this spraying system is to be used in the
future, then a solution needs to be found such that the wind is able to move the
charged droplets away without disturbing the spraying process. One possible
solution is an airfoil that deflects the air flow such that the Taylor cones would be
protected and the wind flows in the volume occupied by charged droplets alone.
Also, similar to the needle based spraying systems, the charged droplets
still are inclined to move towards the charging system. Therefore, after analysis of
the electric field using the same procedure as described in the previous chapter,
steering electrodes can be implemented to facilitate the removal of charged
droplets by the wind.
119
5.4 EWICON with HPMS system1
The HPMS device was implemented in the EWICON system

in a manner similar to the EHDA based spraying systems.
As discussed in the previous chapters, the only spraying
liquid used for the HPMS device(s) was saline water and the charging electrodes
on the device were set on positive potentials as opposed to the negative potentials
needed for EHDA spraying. Also, the fact, that the charged droplets are dispersed
with a high velocity, facilitates the transportation of the droplets by the wind.
Therefore, no steering electrodes were needed in this EWICON configuration.
5.4.1 Single spraying device
In this experiment, one HPMS device has been used. The wind was set at
10 m/s and the flow rate of the water+0.9% NaCl was set at 35 ml/hr. The charging
electrode was set to a potential of 100 V. The charging process can be observed in
Figure 5.9. The potential of the EWICON system rose to 44 kV in 210 seconds. In
Figure 5.9, we can also see that the output power reaches a maximum of 1.63 mW
at 32.5 kV.
1 HPMS device, water+0.9% NaCl
0,100 1,80
0,090 1,60
0,080 1,40
0,070
output power (mW)
1,20
0,060
1,00 Current (µA)
0,050
0,80 Power (mW)
0,040
0,60
0,030
0,020 0,40
0,010 0,20
0,000 0,00
0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00 45,00
Figure 5.9. The indirectly measured charging currents and output power of 1 HPMS
nozzle. The sprayed liquid is water+0.9% NaCl.
1
120
While conducting these experiments, various HPMS devices have been

used. It should be noted that between the various devices, there have been
variations in charging current and output power.
5.4.2 Multiple spraying devices
In this experiment, three HPMS devices have been used. The wind speed
was set at 10 m/s and the total flow rate of the water+0.9% NaCl was set at 105
ml/hr. The charging electrode was set to a potential of 100 V. In Figure 5.10, the
charging current is shown as a function of the output potential of the EWICON
system. There is no measurement data over 50 kV as the spraying system started to
discharge at that moment. In Figure 5.10, we can also see that the output power
reaches a maximum of 5.7 mW at 50 kV or 1.9 mW per HPMS device.
3 HPMS devices, water+0.9% NaCl
0,250 6,00
5,00
0,200
output power (mW)

4,00
0,150
Current (µA)
3,00
Power (mW)
0,100
2,00
0,050
1,00
0,000 0,00
0,00 10,00 20,00 30,00 40,00 50,00 60,00
Figure 5.10. The indirectly measured charging currents and output power of 3 HPMS
nozzles. The sprayed liquid is water+0.9% NaCl.
5.5 Analysis of the experimental results
In this section, we will summarise the charging processes and the output
power of the various configurations of the EWICON system. We will compare
these results with simulations of the equivalent circuit of the EWICON system in
PSpice, of which the procedure will be briefly discussed first.
Subsequently, we will determine whether this output power increases

linearly with the number of spraying nozzles which is of importance with respect to
possible scale-ups of the EWICON system. Next, we will look at the dependence
121
of the output power of some of configurations of the EWICON system on the wind
speed and we will conclude this section by calculating the efficiencies and
EWICON performance indices (EPI).
5.5.1 EWICON circuit equivalent
In order to analyse the behaviour exhibited during the charging

experiments, the EWICON system has been translated to a circuit equivalent. The
configuration in this example is the charging system with the round ring electrodes
spraying the 70%/30% water/ethanol mixture discussed in sections 5.2.1 and 5.3.1.
Both the single needle nozzle and the six needle nozzles version will be analysed.
In this model, which can be found in Figure 5.11, the needle nozzle
spraying the charged droplets has been modelled as a constant current source. The
value of this current source is set at 0.4 μA, which represent the net produced
current produced by the spraying system. In reality, this current will decrease as
the EWICON system potential rises. The source impedance of this current source
has been set to 130 GΩ.
The capacitance of the EWICON system was set at 130 pF and the
resistance of the insulation of the EWICON system was set to 20 TΩ. Finally, a
resistance of 130 GΩ was added to model the electrical load, because this is the
load value at which there is maximum power transfer as can be deduced from
Figure 5.1.
Figure 5.11. The circuit equivalent of the single nozzle EWICON system in PSpice.
The needle nozzle has been modelled as a current source delivering, in this example, a
current of 0.4 μA and an internal source impedance of 130GΩ. The resistance of the
insulators was set at 20 TΩ and the capacitance of the EWICON system was set at 130
pF.
122
First of all, a bias point analysis was performed, i.e. the currents and
potentials were determined while the EWICON system was at a steady-state. The
results have also been displayed in Figure 5.11, where we can see that the
EWICON system potential has reached 24.34 kV and the current through the load
is 187 nA. This agrees reasonably well with the results discussed in section 5.2.1
and Figure 5.1, where the potential is 20 kV and the current is 153 nA.
Secondly, a transient analysis was performed, which gives an indication of

the charging time and the potential of the system. The potential is determined at the
node in Figure 5.11, indicated by the encircled “V”. A plot of the EWICON system
potential as a function of time can be seen in Figure 5.12. We can see that the
potential approaches 24 kV in 35 seconds, after which the system stabilises.
Figure 5.12. A plot of the charging process of the single nozzle EWICON system with
a load of 130 GΩ. The needle nozzle is spraying a current of 0.4 μA. The EWICON
system potential approaches 24 kV after 35 seconds after which the EWICON system
stabilises.
If we now look at the six needle version of the rounded ring electrodes
system, then the model in PSpice becomes as shown in Figure 5.13. Assuming that
the source impedance of each needle nozzle is 130 GΩ, the total source impedance
of the six needles spraying together is 22 GΩ. Also in Figure 5.13, the 20.5 GΩ
load has been connected to the EWICON system similar to the experiment
conducted in section 5.3.1.
123
Figure 5.13. The circuit equivalent of the six nozzles EWICON system in PSpice. The
six needle nozzles have been modelled as one single current source delivering, in this
example, a current of 1.5 μA and an internal source impedance of 22GΩ. The
resistance of the insulators was set at 20 TΩ and the capacitance of the EWICON
system was set at 130 pF. Now, a load of 20.5 GΩ has been attached to the EWICON
system
We can see that the potential reaches 15.74 kV which is in agreement with
the 16 kV measured in section 5.3.1. The current through the 20.5 GΩ load has
been calculated to be 0.78 μA, which is again in agreement with the measurements.
The two other loads of 5.3 and 10.3 GΩ have also been entered in this simulation,
of which the results can be seen in Table 5.1 together with the results of the 20.5
GΩ load.
Table 5.1. Comparison of the measured potentials and currents with the simulated
potentials and currents as a function of the attached electric load.
load measured simulated measured simulated
(GΩ) Uload Uload Iload Iload
(kV) (kV) (μA) (μA)
5.3 5.7 6.4 1.1 1.2
10.3 9.7 10.5 0.9 1.0
20.5 16.0 15.7 0.8 0.8
5.5.2 Charging of the EWICON, output power
The results of the experiments in this chapter have been summarised in the
order of discussion in Table 5.2. The output power Pout has either been given for
the RLOAD or for RPmax. RLOAD is the actual electrical load that has been connected to
124
the EWICON system. RPmax has been calculated as the load at which the maximum
power would be transferred.
Table 5.2. List of the calculated output power. “70w/30e” is the 70%/30%
water/ethanol mixture and “demi water” is demineralised water. UEWICON is the
EWICON system potential and Iout is the current through the electrical load RLOAD.
nozzles UEWICON Iout RLOAD RPmax Pout
# system liquid
no. (kV) (μA) (GΩ) (GΩ) (mW)
1 Delrin 70w/30e 1 2.0 0.1 20 - 0.2
2 demi
Delrin 1 0 0 20 - -
water
3 rounded
70w/30e 1 20 0.15 - 131 3.1
ring
4 demi
cylindrical 1 20 0.018 - 1111 0.4
water
5 demi
rod 8 3.0 - - - -
water
6 rod 70w/30e 8 8.2 0.4 20 - 3.3
7 rod 70w/30e 24 15.0 0.7 20 - 10.5
8 rounded
70w/30e 6 16.0 0.8 20 - 12.5
ring
9 rounded
70w/30e 6 9.7 0.9 10 - 6.4
ring
10 rounded
70w/30e 6 5.7 1.1 5 - 3.3
ring
11 rounded demi
6 15 0.19 79 2.9
ring water
12 demi
cylindrical 3 20 0.14 143 2.7
water
13 tap
cylindrical 3 4.0 0.11 37 0.43
water
14 self-
ethanol 13 2.5 - - - -
adjusting
15 saline
HPMS 1 32.5 0.07 - 650 1.63
water
16 saline
HPMS 3 50 0.15 - 268 5.7
water
125
5.5.3 Analysis of the charging current
• EHDA
In all the EHDA plots of the current as a function of the EWICON system
potential, we can see that the current decreases in gradual manner as expected
because of the polydispersity of the charged droplets. The charged droplets with a
higher electric mobility are attracted back earlier than the charged droplets with a
lower mobility.
• HPMS
In the plots of HPMS experiments, where the current was expected to

decrease sharply at a certain system potential, we can see, however, that this is not
the case. The reason for this behaviour is yet unclear and it was not only seen with
increasing the potential of the holding platform, but also over time. The charging
process may be influenced by local build up of charge in the nozzle and this
requires more investigation.
5.5.4 Analysis of the output power
• EWICON output power in general
Compared to the results of the simulations discussed in chapter 2, the

output power of the EWICON system is lower for all the implementations, both
with EHDA and HPMS, averaging 2.0 mW per nozzle. In chapter 2, it was
calculated that one nozzle spraying at 20 ml/hr theoretically should deliver an
output power of 206 mW.
The main reason for this large difference is that the current associated with
the creation of charged droplets is significantly lower than expected. On average,
the measured current per nozzle was 0.2-0.3 μA, whereas the expected current was
roughly 1.0-1.7 μA assuming that charging efficiencies of 70% of the maximum
given by the Rayleigh limit were reached.
For both EHDA and HPMS, the spraying modes and droplet sizes were not
optimal. With both spraying methods, the droplet diameters were larger than 10
μm. Also, for EHDA, cone-jet spraying for water has not yet been achieved,
leading to lower charging efficiencies. For HPMS, the charging efficiency was in
the order of 3-4% instead of 30-50%.
126
• EHDA
In the case of EHDA based spraying systems, in some cases, the output
power is reasonably proportional to the number of spraying nozzles. In the case of
the rod electrodes charging system, increasing the number of needles from 8 to 24
resulted in an output power increase from 3.3 to 10.5 mW, which is a factor of 3.2,
which can be considered linear.
In the case of the rounded ring electrodes charging system increasing the
number of needles from 1 to 6 resulted in an output power increase from 3.1 to
12.5 mW, which is a factor of 4.0.
In the case of the cylindrical electrodes charging system, increasing the

number of nozzles from 1 to 3 resulted in an output power increase from 0.4 to 2.7
mW, which is a factor of 6.8. The reason that this factor is significantly higher than
3, is that the output power of the single nozzle version is lower. This is due to small
corona discharges which occurred at the two needle tips of the non-spraying
nozzles and thus causing a loss current.
In other multi-nozzle configuration, the single nozzle version could always

be realised by removing the spraying needles. In the case of the configuration with
the cylindrical electrodes this was not possible. Because the nozzles are connected
to the platform, these nozzles will also rise in potential when the EWICON system
is charged, which causes the corona discharge.
• HPMS
In the case of HPMS based spraying systems, in general, linearity of the

output power was found. A single nozzle system yielded on average 1.6 mW,
whereas a three nozzle system yielded 5.6 mW, which is a factor of 3.5.
5.5.5 The output power vs. the wind speed
In order to investigate whether a higher wind speed would result in a

higher output power, an experiment similar to the one described in 4.3.1 was
conducted in which the wind speed was varied from 6 to 12 m/s.
The used set-up in this experiment was a 3 needle nozzle version with the
rounded ring electrodes. The spraying liquid was the 70%/30% water/ethanol
mixture at a flow rate of 20 ml/hr per nozzle. The effective wind surface area was
0.06 m by 0.02 m. The output current through a load of 20.5 GΩ and the output
potential of the EWICON system can be seen in Figure 5.14.
127
current, potential vs. wind speed
1,4 14
1,2 12
1,0 10
potential (kV)
current (µA)
0,8 8
0,6 6
0,4 4 Current (µA)

Potential (kV)
0,2 2
0,0 0
0 2 4 6 8 10 12 14
wind speed (m/s)
Figure 5.14. The output current through and potential over a load of 20.5 GΩ. This
experiment has been conducted with an EHDA based EWICON system consisting of 3
spraying nozzles.
We can see that the current through the load increases with increasing wind
speed, which is consistent with the results of the experiment in section 4.3.1.
Consequently, the EWICON potential rises with the wind speed as well.
In Figure 5.15, the output power has been plotted as a function of the wind
speed together with the maximum convertible power as given by the Betz limit.
Figure 5.15. The output power as a function of the wind speed shown in pink, together
with the fitted curve in yellow. The maximum convertible power from the wind as
given by the Betz limit has also been plotted, shown in red.
128
It should be noted that the order of the Betz power is 100 times higher than
the generated output power.
We can see that the output power rises with increasing wind speed. The
data has been fitted with a cubic polynomial function in MATLAB, see appendix
D, and this yielded the following expression for the output power:
Pout (v w ) = 4.437 ⋅ 10 −3 ⋅ v w3 (5-3)
which has also been included in Figure 5.15. In this fit, the other
coefficients have been set to zero. We can conclude that the output power of the
EWICON system relates to the 3rd power of the wind speed. It has to be noted,
though, that the EWICON system can convert the wind power to electrical power
only if sufficient charged droplet are available to be moved by the wind.
5.5.6 Efficiencies of the spraying systems, EPI
In section 2.3, the efficiency, η, and the EWICON Performance Index, EPI
were defined and, for all the systems discussed in Table 5.2, these have been
calculated for as shown in Table 5.3. The wind surface area, A, has been
determined by visual inspections of the sprayed droplets as they were sprayed away
from the nozzles. Together with the wind speed, the maximum recoverable power,
Pmax, wind, could be calculated with (2-2).
The electrical power, Pelectrical, for EHDA based spraying was estimated by
observing that there was no measurable current (< 0.05 μA) flowing back from the
electrodes to the DC source supplies. At any given time during these experiments,
the sum of the potentials applied to each of these electrodes was never higher than
20 kV. Therefore, the upper limit of the required electrical power was estimated to
be 0.1 mW.
The power required to pump the spraying liquid or the mechanical power,
Pmechanical, has been calculated using (4-4) for a height of 2 metres.
129
Table 5.3. The efficiencies and the EWICON Performance Indices of all the discussed
configurations listed in Table 5.2. The grey areas are the experiments that did not
yield an output power.
Pmax,
Pout Pelectrical Pmechanical Pin Q A η
# EPI
(mW) wind
(mW) (mW) (mW) (l/h) (cm2) (%)
(mW)
EHDA systems
1 0.2 33 0.1 0.109 33.209 0.02 4 0.6 2.2
2
3 3.1 33 0.1 0.109 33.209 0.02 4 1.17 3.7
4 0.4 33 0.1 0.109 33.209 0.02 4 0.15 1.9
5
6 3.3 265 0.1 0.872 265.972 0.16 32 0.27 1.2
7 10.5 794 0.1 2.616 796.716 0.48 96 0.29 0.8
8 12.5 199 0.1 0.654 199.754 0.12 24 0.79 2.5
9 6.4 199 0.1 0.654 199.754 0.12 24 0.40 2.0
10 3.3 199 0.1 0.654 199.754 0.12 24 0.21 1.4
11 2.9 199 0.1 0.654 199.754 0.12 24 0.18 1.3
12 2.7 99 0.1 0.327 99.427 0.06 12 0.34 2.1
13 0.43 99 0.1 0.327 99.427 0.06 12 0.05 0.5
14
HPMS systems
15 1.63 345 0.0 1.6 345.127 0.035 9 0.47 2.4
16 5.7 1034 0.0 5.1 1039.1 0.105 27 0.55 2.0
The first observation that can be made is that, with exception of

configuration #1, all configurations, both EHDA and HPMS, yielded a higher
output power than the required electrical and mechanical power. Secondly, if we
look at Pin which is comprised of Pmax,wind, Pelectrical and Pmechanical, then the main
factor in Pin is the maximum recoverable power from the wind, Pmax,wind, and thus
the determining factor in the efficiencies.
With regard to spraying liquids for EHDA based spraying systems, the
efficiencies are higher when water/ethanol mixtures are sprayed than tap or
demineralised water. Also, if we look at the efficiencies of the EHDA based
130
spraying systems, then the rounded ring electrodes yield the highest efficiencies.
Based on the EPI, configurations #3 and #8 are the best performers.
If we compare the EPIs of EHDA and HPMS, then it can be concluded that
there is no significant difference between them. Therefore, at this point, it can be
concluded that with respect to the EWICON system, both spraying methods are
equally suitable.
5.6 Conclusions
Overall, we can conclude that the EWICON system can be charged up to a

potential by allowing the wind to move charged droplets against an electric field.
This was possible using both EHDA and HPMS as the spraying method. It was
also possible to deliver output power to an electrical load.
In most cases, the output power of the EWICON system was higher than
the required electrical and mechanical power. However, if the output power is
compared to the maximum recoverable power in the wind, the conversion
efficiency is very low and in most case less than a percent. This is due to two
reasons. Firstly, the current associated with the droplet creation process per nozzle
is lower than expected, which has been observed with both EHDA and HPMS.
Secondly, for EHDA, this current does not scale proportionally with the number of
nozzles.
5.7 Scaling up the EWICON

5.7.1 General considerations
In chapter 2, a few words were spent on the scaling up of the EWICON

system. If 900 nozzles were positioned in one square metre, then an output power
of 102 W could be achieved. Here, we assumed that we could spray 10 μm droplets
with 70% of the maximum charge as given by the Rayleigh limit.
In such a scenario, a 100 kilowatt version of the EWICON system could be

achieved by placing roughly 106 nozzles with 2 cm spacing in a frame sized 20
metres by 20 metres. The current sprayed by the nozzles would then be 1.7 ampere
and the EWICON system operating potential would then be roughly 60 kV.
Concerning public safety, the following can be said. If such a structure is

placed at a height of e.g. 30 metres, the electric field at ground level would already
be lower than the 600 V/m, which is the electric field strength at which no harmful
effects on humans have been observed, [42]. A current produced by the EWICON
131
system will produce a magnetic field and, since this current is a DC current, the
magnetic field will be static. According to [43], even high static magnetic fields
have no proven harmful effects on humans, except in indirect cases.
5.7.2 Extrapolation based on the conducted experiments
Considering the results stated in Table 5.2 and Table 5.3, we can conclude
that the considerations in the previous section are premature. If we take the best
multi-nozzle system result, which is the 6 nozzles with the rounded ring electrodes
configuration spraying the water/ethanol mixture, then the output power per nozzle
is 2.1 mW.
A 1 kilowatt version using this configuration would require 4.8·105

spraying nozzles, roughly 700 by 700 nozzles or 14 metres by 14 metres if we
assume 2 cm spacing. This means that spraying nozzles need to be more closely
packed together before a kilowatt version will be feasible from an economical point
of view.
For EHDA, achieving this with spraying needles will be a serious

challenge and, therefore, the self-adjusting nozzles concept should be further
investigated as an option for charged droplet generation. It was already possible to
create 13 Taylor cones in the same area where normally only 6 to 8 spraying
needles could be fitted.
For HPMS, even with the lower than expected charging efficiencies, the
increase in produced current should be feasible. At this point in time, the devices
are equipped with 34 pores. The current technology allows for hundreds of pores to
be manufactured on the same area. This should increase the current density output
and thus provide a good option for a scaled-up EWICON.
132
Chapter 6 – Conclusions and recommendations
6
Conclusions and recommendations
“All truths are easy to understand once they are discovered; the point is to discover
them.”
– Galileo Galilei
In this thesis, the EWICON concept has been introduced as an interesting

future alternative to the conventional wind turbine generator. In the first few
chapters, the theoretical foundation was provided on which, in the subsequent
chapters, the practical aspects could be built, such as droplet creation and charging
methods or the choice of spraying liquids.
Firstly, in section 6.1, we will recap the conclusions for each chapter, after
which a number of general conclusions will be drawn with regard to the practical
feasibility of the EWICON concept.
Lastly, in section 6.2, we will give some recommendations on future work
and steps that need to be undertaken in order to solve crucial scientific obstacles
and thus increase the commercial viability of an industrial implementation of the
EWICON concept.
6.1 Conclusions
6.1.1 Conclusions per chapter
In chapter 1, we concluded that the Type B implementation of the

EWICON concept, i.e. the implementation with a charging system isolated from
earth without a collector unit, is best suited for commercial exploitation. The fact
that no separate collector unit is needed means that product costs are lower. Also,
the wind only has to move the charged droplets to earth instead of towards a certain
fixed point. Otherwise, this would require a relatively large unit to take in all the
droplets or wind-direction following system.
In chapter 2, we concluded, from a theoretical point of view using an

analytical model with a small number of charged droplets, that it is possible to
133
transport these droplets by the wind against the electric field. We also concluded
that the increase of electrostatic energy is sufficiently high to make use of the
potential energy contained in the wind, provided that a high density of droplet
creation units could be achieved.
Using this theoretical approach, parameters, like e.g. the minimum required
wind speed as a function of droplet diameter, charge and the existing electric field
have been determined. If the electric field is not higher than 104 V/m, then the
minimum wind speed for droplets diameters varying from 0.1 to 100 μm is less
than 3.3 m/s (2 on the scale of Beaufort). These droplets are assumed to be highly
charged, which, in this case, is 70% of the maximum charge allowed on a droplet.
In chapter 3, we concluded that out of a number of droplet creation and

charging methods, at this moment, two viable methods exist, each having their own
merits and drawbacks.
The first method is electrohydrodynamic atomisation (EHDA) and two
important factors have been found for successful operation. Firstly, the surface
tension of the liquid should preferably be low, in the order of 40·10-3 N/m.
Secondly, the electric field, responsible for the creation and charging should be
focussed on the liquid exiting the nozzle.
The main advantage of EHDA is that it requires little input power
compared to the other methods available. However, while the results of using water
as the spraying liquid have improved significantly during the course of this thesis
work, it still has not been possible to make full use of the other main advantage of
EHDA, namely the production of highly charged monodisperse droplets.
The second method is high pressure monodisperse spraying (HPMS). The
main advantage of HPMS is that it is possible to use (saline) water as the spraying
liquid, which is positive with respect to availability and environmental issues. The
second advantage is the fact that the droplets produced by HPMS are
monodisperse, which facilitates the optimisation of the EWICON system.
However, the required power is higher compared to the EHDA method, because a
high pressure is needed to create the droplets.
In chapter 4, we concluded that it was possible to design the EWICON

system in such a way, that the charged droplets were actually removed by the wind,
thus creating a charging current. Using both EHDA and HPMS methods, several
multiple-nozzle spraying systems were constructed and tested, both showing
varying degrees of linearity with respect to the output current.
In chapter 5, we concluded that the transportation of the charged droplets

by the wind resulted in an output current that could charge the EWICON system up
to a potential, depending on the wind speed and the charging current. It was
possible to connect the EWICON to an electrical load which resulted in an output
power. As expected, this output power increased with the wind speed. With both
134
spraying methods, it was possible to achieve a positive conversion efficiency, even

though this efficiency was a few percent.
Based on these results, the output power per wind surface area needs to be
increased by a factor of 50 in order to make an EWICON system comparable to a
conventional wind turbine, thus making a large scale EWICON feasible.
6.1.2 Overall conclusions on the EWICON concept
In general, the EWICON concept has been proven to be able to achieve a

net gain in terms of output power depending on the used methods and more
importantly, the used spraying liquid. Using mixtures of water and ethanol in
combination with the EHDA method proved that it was possible to generate high
droplet currents which resulted in positive conversion efficiencies. However, using
tap, salt or demineralised water, yielded lower conversion efficiencies, even when
considerable amount of effort was put into the design of the charging system.
Therefore, the EHDA method needs to be optimised to spray water with high
currents, in order to develop a commercially viable wind energy converter based on
the EWICON concept.
With respect to the HPMS method, also a positive gain was achieved.
However, at this point in time, the spraying devices that were available did not
allow for a high number of nozzles with respect to the wind surface area. Also, the
charging process was found to be insufficient as the charge on the droplets was 3-
4% of the maximum as given by the Rayleigh limit. Therefore, while again the
EWICON concept equipped with HPMS proves to be successful in the laboratory
environment, the charging process needs to be upgraded in order for a
commercially viable product to be produced.
Looking at the considerations made on the subject of up scaling, we can

conclude, that it is reasonably feasible to increase the current production per wind
surface area with both methods, which means that a commercially viable product is
still a realistic goal.
The EHDA spraying method still has several unexplored implementations,
like the optimisation of the cylindrical water spraying electrodes or the self-
adjusting multi-nozzles set-up. Both concepts have the potential to increase the
output current by optimising the spraying process or by either increasing the rate of
charged droplets.
The HPMS devices can be designed in such a way that a significantly
higher number of pores can be applied than the 34 pores currently present on the
device. New experiments will have to prove that the output current scales linearly
with the increase of pores. Moreover, earlier testing has already proved that
increasing the number of spraying devices results in a linear increase of the output
current.
135
6.2 Recommendations
The main cause of the lower output power of the EWICON system with
EHDA is the fact that the total produced current is lower than what is expected
based on the proportionality of the spraying nozzles. Therefore, research should be
focussed on the processes that take place on a fundamental level when multiple
spraying nozzles are electrically connected. The effect of the different materials at
the liquid/nozzle interface on the mobility of the charged carriers should be
investigated.
Preferably, these experiments should be conducted in a wind tunnel that
has sufficiently large dimensions, so that a high number of nozzles can be placed
together. In this way, any fringe effects by the outer nozzles will have a negligible
impact on the total output current.
With respect to the EHDA spraying method, the use of water should be
thoroughly investigated. In the field of electrospray, the spraying of water does not
have a high priority compared to the spraying of medicine or coating material.
However, in order for the EWICON system to be commercially successful and
environmentally acceptable, the electrospraying of tap and salt water is a
prerequisite. The cylindrical nozzles described in chapter 3 can act as a base point
for further research in which also the high conductivity of tap and salt water has to
be taken into account.
With respect to the HPMS method, spraying devices with a higher

concentration of pores should be developed. Also, keeping in mind that, because of
the affinity for saline water, HPMS devices are a candidate for an EWICON at sea,
filtering techniques should be developed as to prevent the clogging of the pores.
On the matter of alternative methods for liquid supply to a certain height,

condensation of water in greenhouses is an interesting phenomenon. A greenhouse
area of 10000 m2, which is a prevalent size in the Netherlands, could potentially
produce 2000 litres per hour of water. In this way, the spraying liquid could be
transported to the EWICON operating height without the use of pumps.
A substantial part of the losses in the EWICON system are due to the
inability of the charged droplets to reach earth. Calculations have shown that the
electric field generated by the whole EWICON system should not be higher than
104-105 V/m anywhere along the trajectory of the charged droplets. Any new
charging system should be modelled in a simulation software package to ascertain
that this is the case.
136
Appendices
137
Appendix A. Betz’ law
In this appendix, we will derive the maximum recoverable power from the
wind in a manner analogous to the derivation done by Betz [16] and Lanchester
[17].
First of all, we will define the parameters. The wind velocity far upstream of
the converter is v1 (vw in chapter 2), the wind velocity at the converter is v and the
wind velocity far downstream of the converter is v2. This has been illustrated in
Figure A.1.
Figure A.1. Illustration of the Betz limit.
At these points, the wind flows through surface areas A1, A and A2. If we
look at the mass of air, m, that passes through an area A in a time dt, then
m = ρ ⋅ A ⋅ dx (A-1)
in which ρ is the density of air and dx is the distance travelled by the wind
with velocity, thus
139
dx = v ⋅ dt (A-2)
The kinetic energy associated with this flow of air is
Ek = 12 ⋅ m ⋅ v 2 (A-3)
which with equation (A-1) can be written as
Ek = 12 ⋅ ρ ⋅ A ⋅ v 3 ⋅ dt (A-4)
which means that the power associated with the wind is
P = 12 ⋅ ρ ⋅ A ⋅ v 3 (A-5)
which is equation (2.1).
In order to get to equation (2.2), we first need to determine the wind

velocity, v, at the convertor as a function of the wind velocity up and downstream
of the converter, v1 and v2 respectively. We do this by considering the continuity
equation
ρ1 ⋅ A1 ⋅ v1 = ρ ⋅ A ⋅ v = ρ 2 ⋅ A2 ⋅ v2 (A-6)
which means that the mass flow rate is constant. The momentum flow rate is
the mass flow rate times the velocity or ρ·A·v2.The change in the momentum flow
rate flowing from A1 to A2 is equal to the force, F, acting on the converter. The
power, P, is then F·v, so that we can write for the power extracted from the wind
(
Pextracted = ρ ⋅ A1 ⋅ v12 − ρ ⋅ A2 ⋅ v22 ⋅ v) (A-7)
We can also write for the extracted power, the change in power associated
with the wind as it passes through the converter
Pextracted = 12 ⋅ ρ1 ⋅ A1 ⋅ v13 − 12 ⋅ ρ 2 ⋅ A2 ⋅ v23 (A-8)
If we equate (A-7) and (A-8), we find by using the continuity equation (6)
that the wind velocity at the converter, v, is the average of v1 and v2
v = 12 ⋅ (v1 + v2 ) (A-9)
140
If we use this result in (A-7) with the continuity equation (A-6), then we find
that the power extracted by the wind is
(
Pextracted = 14 ⋅ ρ ⋅ A ⋅ v13 − v23 + v12 v2 − v1v22 ) (A-10)
The maximum of this function can be found by
dPmax
=0 (A-11)
dv2
which gives us
v2 = 13 ⋅ v1 (A-12)
If we insert this in equation (A-10), then we find for the maximum

recoverable power from the wind that
Pmax = 8
27 ⋅ ρ ⋅ A ⋅ v13 (A-13)
which is equation (2.2) where vw is v1.
141
Appendix B. Evaporation and relaxation times

of droplets
In this appendix, based on [18], we will give the rate of evaporation, which
is the rate of change of the droplet diameter, and the droplet lifetime, which is the
time required for a droplet to evaporate completely. We will briefly indicate in
what way the wind speed and the relative humidity, also called the saturation ratio
SR, influence the droplet lifetime, t. We will also discuss how to determine the
relaxation times of droplets.
The rate of evaporation for droplets with a droplet diameter d1 > 0.1 μm is
d (d d ) 4 Dv M ⎛ p∞ p d ⎞
= ⎜⎜ − ⎟⎟ (B-1)
dt Rρ d d d ⎝ T∞ Td ⎠
in which dd is the droplet diameter, Dv is the diffusion coefficient, M is the

molar mass, R is the universal gas constant, ρd is the density of the liquid, p∞ is the
partial pressure surrounding the droplet, pd is the partial pressure at the droplet
surface, T∞ is the temperature of the surrounding vapour and Td is the temperature
at the droplet surface.
By integrating (B-1) from d1 to 0 for dd, we can determine the droplet

lifetime t
Rρ d d12
t= (B-2)
⎛p p ⎞
8 Dv M ⎜⎜ d − ∞ ⎟⎟
⎝ Td T∞ ⎠
which according to [18] is accurate for calculating the droplet lifetimes of

droplets with a diameter from 10 to 50 μm, because these droplets are larger than
1.0 μm most of their lifetime. Above 1.0 μm, the droplet life time is proportional to
the square of the droplet diameter. In a dry environment, the partial pressure of the
surrounding vapour, p∞, is negligible.
143
For Td, pd and Dv empirical relations exist. The following relation exists
between Td and T∞
Td − T∞ =
(6.65 + 0.345 ⋅ T∞ )
+ 0.0031 ⋅ T∞2 (S R − 1)
(B-3)
1 + (0.082 + 0.00782 ⋅ T∞ ) ⋅ S R
which is valid for ambient temperatures, T∞, of 0–40 °C and saturation

ratios, SR, of 0–5. The following relation exists for pd
⎛ 4060 ⎞
pd = exp⎜⎜16.7 − ⎟ (B-4)
⎝ Td − 37 ⎟⎠
which is valid for ambient temperatures of 0–100 °C. The following relation
exists for the diffusion coefficient of water vapour in air, Dv
Dv = 21.2 ⋅ 10 −6 (1 + 0.0071 ⋅ T ) (B-5)
which is valid for ambient temperatures of 0–45 °C. Using these equations,
the evaporation times can be calculated for a water droplet in dry static air.
For example, a droplet with a diameter of 20 μm would take 0.31 seconds to

fully evaporate in dry static air at 20 °C. In order to take into the account the effect
of wind, we have to calculate the relaxation time, τ, of the droplet. This is the time
that a droplet needs to adjust to a new set of conditions of forces, which according
to [18] can be calculated as follows
ρ d d d2
τ= (B-6)
18η
in which η is the viscosity of air. If we calculate the relaxation times and

compare them to the evaporation times of the droplets, then, in general, we see that
the relaxation time is lower than 0.5% of the evaporation time. In other words, on
the scale of the evaporation time, the droplet almost immediately acquires the
prevailing wind speed and, therefore, expression (B-2) can be used to estimate the
evaporation time.
144
Another issue concerning the evaporation of droplets is the relative humidity

or saturation ratio. The evaporation times have been calculated for a single droplet
in dry air where p∞ is assumed to be zero. However, in the case of the EWICON,
the droplets are sprayed at a rate of 106 per second in a small volume. Therefore,
we can assume that the relative humidity in this mist of droplets is between 80%
and 100%.
In Figure B.1, taken out of [18], (B-2) has been calculated for several
relative humidities. We can see that at a relative humidity of 100%, the droplet
lifetime is in the order of 10 seconds for droplets larger than 5 μm.
Figure B.1. The droplet lifetimes as a function of the droplet diameter at relative
humidities of 0%, 50% and 100%. We can see that at a relative humidity of 100%, the
droplet lifetime is in the order of 10 seconds for droplets larger than 5 μm.
145
Appendix C. Numerical modelling in MATLAB
In this appendix, the listing of the MATLAB program used in chapter 2 will
be given. The numerical method used to calculate the positions, velocities and
accelerations of the charged droplets is Velocity-Verlet.
% M-File: TenDroplets
%
% This programs will simulate 10 charged droplets in an electric field
% under the influence of wind and gravity
%
% D.Djairam 04/04/2008
%
% Variables
% eo free space permittivity (F/m)
% g gravity acceleration (m/s^2)
% rho_droplet density of liquid (kg/m^3)
% rho_air density of air (kg/m^3)
% gamma surface tension liquid (N/m)
% eta dynamic viscosity
% d1..d10 droplet diameter (m)
% rayl1.. rayl10 charge on droplet in terms of the Rayleigh limit
% vw wind speed (m/s)
% Eex external electric field (V/m)
% totaltime total simulation time (s)
% dt time step (s)
% dtsq time step squared (s)
% qm1..qm10 maximum charge on the droplets (C)
% q1..q10 charge on the droplets (C)
% x,v 1..10 position and velocity for all ten droplets
% m1..m10 mass of the droplets (kg)
% Fg gravity force (N)
% Fb buoyancy (N)
% Fw wind force due to drag
% Fex electric force due to external field
% Fqq electric forces due to other charged droplets
% Initialize variables
eo=8.854e-12;
g=-9.81;
K=1/(4*pi*eo);
rho_droplet=0.998e3;
rho_air=1.225;
gamma=7.28e-2;
eta=17.8e-6;
d1=10e-6;d2=10e-6;d3=10e-6;d4=10e-6;d5=10e-6;
d6=10e-6;d7=10e-6;d8=10e-6;d9=10e-6;d10=10e-6;
rayl1=0.7;rayl2=0.7;rayl3=0.7;rayl4=0.7;rayl5=0.7;
rayl6=0.7;rayl7=0.7;rayl8=0.7;rayl9=0.7;rayl10=0.7;
vw=[10 0 0];
Eex=-80000;
totaltime=1;
dt=0.00001;
dtsq=dt*dt;
N=int32(totaltime/dt); % number of calculations
147
% Preallocate time and positions matrices for end result storage

time=zeros(N+1,1);
x1pos=zeros(N+1,3);x2pos=zeros(N+1,3);x3pos=zeros(N+1,3);x4pos=zeros(N+1,3);
x5pos=zeros(N+1,3);x6pos=zeros(N+1,3);x7pos=zeros(N+1,3);x8pos=zeros(N+1,3);
x9pos=zeros(N+1,3);x10pos=zeros(N+1,3);
% Calculate the charges on the droplets

qm1=8*pi*sqrt(eo*gamma*(d1/2)^3);q1=rayl1*qm1;
% Calculate the mass of the droplets

volume1=(4/3)*pi*(d1/2)^3;m1=volume1*rho_droplet;
% Assigning starting position and velocity of the droplets

x1=[1 5.99 4.99];v1=[0 0 0];
x2=[1 6.00 4.99];v2=[0 0 0];
x3=[1 6.01 4.99];v3=[0 0 0];
x4=[1 5.99 5.01];v4=[0 0 0];

x5=[1 6.00 5.01];v5=[0 0 0];
x6=[1 6.01 5.01];v6=[0 0 0];
x7=[1 5.99 5]; v7 =[0 0 0];

x8=[1 5.995 5];v8 =[0 0 0];
x9=[1 6.005 5];v9 =[0 0 0];
x10=[1 6.01 5];v10=[0 0 0];
time(1,1)=0;
x1pos(1,1:3)=x1;x2pos(1,1:3)=x2;x3pos(1,1:3)=x3;x4pos(1,1:3)=x4;x5pos(1,1:3)=x5;
x6pos(1,1:3)=x6;x7pos(1,1:3)=x7;x8pos(1,1:3)=x8;x9pos(1,1:3)=x9;x10pos(1,1:3)=x10;
% Start with work done on droplet set to zero

W1=0;W2=0;W3=0;W4=0;W5=0;W6=0;W7=0;W8=0;W9=0;W10=0;
% Start calculation
% from 1 until N
for i = 1:N
% -------------------------------------------
% Calculation distances and unit vectors
r12=magvector(x1-x2);u12=unitvector(x1-x2);r13=magvector(x1-x3);u13=unitvector(x1-x3);
r110=magvector(x1-x10);u110=unitvector(x1-x10);
148
% Calculation forces on droplets: Stokes' force, electric force from

% external field, electric force from other charged droplets and gravity
% Stokes' force
Fw1=[6*pi*eta*(vw(1)-v1(1))*(d1/2) 6*pi*eta*(vw(2)-v1(2))*(d1/2) 6*pi*eta*(vw(3)-
v1(3))*(d1/2) ];
v2(3))*(d2/2) ];
v3(3))*(d3/2) ];
v4(3))*(d4/2) ];
v5(3))*(d5/2) ];
v6(3))*(d6/2) ];
149

v7(3))*(d7/2) ];
v8(3))*(d8/2) ];
v9(3))*(d9/2) ];
v10(3))*(d10/2) ];
% The electric force from external E-field.

% We assume that the E-field drops off
% with the square of the x-distance and that the source of the E-field is
% at the plane x=0
Fex1=[q1*sign(x1(1))*Eex/x1(1)^2 0 0];Fex2=[q2*sign(x2(1))*Eex/x2(1)^2 0 0];
% mutual electric forces from charged droplets
Fq1q2=K*q1*q2/r12^2*u12;Fq1q3=K*q1*q3/r13^2*u13;
Fq1q10=K*q1*q10/r110^2*u110;
Fq2q10=K*q2*q10/r210^2*u210;
Fq3q10=K*q3*q10/r310^2*u310;
Fq4q10=K*q4*q10/r410^2*u410;
Fq5q10=K*q5*q10/r510^2*u510;
Fq6q10=K*q6*q10/r610^2*u610;
Fq7q10=K*q7*q10/r710^2*u710;
Fq8q10=K*q8*q10/r810^2*u810;
150
Fq9q10=K*q9*q10/r910^2*u910;
Fq10q9=K*q10*q9/r109^2*u109;
% gravity forces
Fg1=[0 0 m1*g];Fg2=[0 0 m2*g];Fg3=[0 0 m3*g];Fg4=[0 0 m4*g];
Fg9=[0 0 m9*g];Fg10=[0 0 m10*g];
% buoyancy
Fb1=[0 0 -volume1*rho_air*g];Fb2=[0 0 -volume2*rho_air*g];Fb3=[0 0 -volume3*rho_air*g];
Fb10=[0 0 -volume10*rho_air*g];
% sum of all forces on droplets

SF1=Fw1+Fex1+Fq1q2+Fq1q3+Fq1q4+Fq1q5+Fq1q6+Fq1q7+Fq1q8+Fq1q9+Fq1q10+Fg1+Fb1;
SF10=Fw10+Fex10+Fq10q1+Fq10q2+Fq10q3+Fq10q4+Fq10q5+Fq10q6+Fq10q7+Fq10q8+Fq10q9+Fg10+Fb1
0;
% -------------------------------------------
% Velocity Verlet algorithm in four steps

% first: calculate new position at whole time step
x1=x1+dt*v1+dtsq/2*SF1/m1;
x10=x10+dt*v10+dtsq/2*SF10/m10;
% second: calculate new speed at half time step

v1=v1+dt/2*SF1/m1;
v2=v2+dt/2*SF2/m2;
v3=v3+dt/2*SF3/m3;
v4=v4+dt/2*SF4/m4;
v5=v5+dt/2*SF5/m5;
v6=v6+dt/2*SF6/m6;
v7=v7+dt/2*SF7/m7;
v8=v8+dt/2*SF8/m8;
v9=v9+dt/2*SF9/m9;
v10=v10+dt/2*SF10/m10;
%third: calculate forces with new position and half-step velocity

% -------------------------------------------
% Calculation distances and unit vectors 2
151
% Calculation forces on droplets: Stokes' force, electric force from
% external field, electric force from other charged droplets and gravity
% Stokes' force 2
v1(3))*(d1/2) ];
v2(3))*(d2/2) ];
v3(3))*(d3/2) ];
v4(3))*(d4/2) ];
152

v5(3))*(d5/2) ];
v6(3))*(d6/2) ];
v7(3))*(d7/2) ];
v8(3))*(d8/2) ];
v9(3))*(d9/2) ];
v10(3))*(d10/2) ];
% The electric force from external E-field 2.
% We assume that the E-field drops off
% with the square of the x-distance and that the source of the E-field is
% at the plane x=0
% mutual electric forces from charged droplets 2
Fq1q10=K*q1*q10/r110^2*u110;
Fq2q10=K*q2*q10/r210^2*u210;
Fq3q10=K*q3*q10/r310^2*u310;
Fq4q10=K*q4*q10/r410^2*u410;
Fq5q10=K*q5*q10/r510^2*u510;
Fq6q10=K*q6*q10/r610^2*u610;
Fq7q10=K*q7*q10/r710^2*u710;
Fq8q10=K*q8*q10/r810^2*u810;
153
Fq9q10=K*q9*q10/r910^2*u910;
Fq10q9=K*q10*q9/r109^2*u109;
% gravity forces 2
Fg9=[0 0 m9*g];Fg10=[0 0 m10*g];
% buoyancy 2
Fb10=[0 0 -volume10*rho_air*g];
% sum of all forces on droplets 2
SF10=Fw10+Fex10+Fq10q1+Fq10q2+Fq10q3+Fq10q4+Fq10q5+Fq10q6+Fq10q7+Fq10q8+Fq10q9+Fg10+Fb1
0;
SA1=SF1-Fw1;SA2=SF2-Fw2;SA3=SF3-Fw3;SA4=SF4-Fw4;SA5=SF5-Fw5;
SA6=SF6-Fw6;SA7=SF7-Fw7;SA8=SF8-Fw8;SA9=SF9-Fw9;SA10=SF10-Fw10;
% -------------------------------------------
% fourth: calculate new speed at whole time step

v1=v1+(dt/2)*(SF1/m1);v2=v2+(dt/2)*(SF2/m2);v3=v3+(dt/2)*(SF3/m3);
v10=v10+(dt/2)*(SF10/m10);
% write position and time to file or visualizer

time(i+1,1)=i*dt;
x1pos(i+1,1:3)=x1;x2pos(i+1,1:3)=x2;x3pos(i+1,1:3)=x3;x4pos(i+1,1:3)=x4;
x5pos(i+1,1:3)=x5;x6pos(i+1,1:3)=x6;x7pos(i+1,1:3)=x7;x8pos(i+1,1:3)=x8;
x9pos(i+1,1:3)=x9;x10pos(i+1,1:3)=x10;
% calculate the work dW = F.dl using the SF per droplet
W1=W1+SA1(1)*(x1pos(i+1,1)-x1pos(i,1))+SA1(2)*(x1pos(i+1,2)-
x1pos(i,2))+SA1(3)*(x1pos(i+1,3)-x1pos(i,3));
154
% repeat N times
end
% plot the trajectories

t = 1:1000:N;
plot3(x1pos(t,1),x1pos(t,2),x1pos(t,3))
hold on
xlabel('X Axis')
ylabel('Y Axis')
zlabel('Z Axis')
grid on
hold off
155
Appendix D. Statistical analysis in MATLAB

In this appendix, the results of the statistical analysis that provided the fitting
of the experimental data discussed in chapter 5 will be given. This will concern the
single and six nozzle versions of the rounded ring electrode charging system.
Also, the results of the fitting of the output power as function of the wind
speed will be provided.
D.1 Rounded ring electrode – single nozzle
Firstly, the results of the measurements of a single nozzle version were fitted
with a decreasing exponential curve of which the result can be seen in Figure D.1
together with the 95% confidence bounds. The exponential curve is given by
I c (U EWICON ) = 0.501 ⋅ 10 −6 exp(−62.31 ⋅ 10 −6 ⋅ U EWICON ) (D-1)
in which Ic is the charging current and UEWICON is the EWICON system

potential.
Rounded ring electrode, 1 nozzle, 70%/30% water/ethanol: measurement + fitted curve

550
I vs. U
500 1 nozzle fit
Pred bnds (1 nozzle fit)
450
400
350
charging current (nA)
300
250
200
150
100
50
0
0 5 10 15 20 25 30 35 40 45 50
Figure D.1. The measurement of a single nozzle version together with the fitted curve
and the 95% confidence bounds.
157
In this case, the coefficient of determination, R2, was found to be 0.9824.
D.2 Rounded ring electrodes – six nozzles
Secondly, the results of the measurements of a six nozzle version were fitted
with a decreasing exponential curve of which the result can be seen in Figure D.2
together with the 95% confidence bounds. The exponential curve is given by
I c (U EWICON ) = 1.581 ⋅ 10 −6 exp(−58.85 ⋅ 10 −6 ⋅ U EWICON ) (D-2)
Rounded ring electrode, 6 nozzles, 70%/30% water/ethanol: measurement + fitted curve

2000
I2 vs. U2
6 nozzles fit
1800 Pred bnds (6 nozzles fit)
1600
1400
charging current (nA)
1200
1000
800
600
400
200
0
0 5 10 15 20 25 30 35 40 45 50
Figure D.2. The measurement of a six nozzle version together with the fitted curve and
the 95% confidence bounds.
158
D.3 Output power vs. wind speed
Finally, the output power, Pout, has been measured as a function of the wind
speed, vw, and the results have been fitted with a cubic polynomial curve
y ( x ) = c0 x 3 + c1 x 2 + c2 x + c3 (D-3)
where all the coefficients except c0 have been set to zero. This resulted in an
expression for the output power Pout
Pout (v w ) = 4.437 ⋅ 10 −3 ⋅ v w3 (D-4)
The result of the fitting can be seen in Figure D.3.
Rounded ring electrode, 3 nozzles, 70%/30% water/ethanol: output power vs. wind speed + fitted curve
10
7
output power (mW)
Pout vs. wind

cubic fit
2 Pred bnds (cubic fit)
0
0 2 4 6 8 10 12 14
wind speed (m/s)
Figure D.3. The output power measurement of a three nozzle EWICON system as a
function of the wind speed together with the fitted curve and the 95% confidence
bounds
159
Appendix E. LORENTZ
LORENTZ is a computer aided engineering (CAE) software by Integrated

Engineering Software that has been used in this research to model and analyse the
EWICON system.
In addition to calculating the electric fields generated by the electrodes

present in the EWICON system, LORENTZ is also able to compute the trajectories
of charged particles in the presence of electric and/or magnetic fields. During the
calculation of the trajectories of the charged particles, LORENTZ also takes into
account mechanical forces such as gravity and/or wind forces.
The numerical computational method that is employed by LORENTZ to

solve the electric field is the boundary element method (BEM). The main
advantage of this method over finite element method (FEM) is that only a mesh of
the surfaces of the electrodes is required in order to provide a complete solution.
This simplifies the modelling effort and it can reduce the calculation times when
the volume to surface ratio of the electrodes is large. One disadvantage of BEM is
that it is only applicable to certain linear problems, whereas FEM can handle
nonlinear problems well.
Most electrostatic problems discussed in this thesis, however, are assumed

to be linear and have been solved using BEM. The problems in which the
trajectories of charged particles are involved the solution are obtained by using a
hybrid technique of boundary and volume meshes.
The reader is referred to the website of Integrated Engineering Software,

www.integratedsoft.com, for more detail on LORENTZ and the numerical
techniques that are employed for solving models.
161
List of symbols and abbreviations
symbol unit description

A m2 – Wind surface area
Cdroplet F – Capacitance between droplet and charging
electrode
Cc - – Cunningham slip correction factor
CD - – Drag coefficient
dd m – Droplet diameter
E V·m-1 – Electric field
ε0 C2·s2·kg-1·m-3 – Permittivity of vacuum
εr - – Liquid permittivity
F N – Force
γ N·m-1 – Surface tension
h m – Height of the EWICON system
ηEWICON - – EWICON efficiency
ηa Pa·s – Dynamic/Absolute viscosity
I A – Current
K S·m-1 – Conductivity
μ Pa·s – Dynamic/Absolute viscosity
μe m2·V-1·s-1 – Electrical mobility
μm s·kg-1 – Mechanical mobility
μg - – Geometric mean
Pin W – Input power
Pmax W – Maximum recoverable wind
Pout W – EWICON output power
Pw W – Power associated with the wind
Q m3·s-1 – Liquid flow rate
q C – Charge
qmax C – Maximum charge on a droplet
Re - – Reynolds number
ρ kg·m-3 – Density
ρa kg·m-3 – Density of air
σ - – Standard deviation
σg - – Geometric standard deviation
vd m·s-1 – Droplet velocity
163
vT m·s-1 – Terminal electrostatic velocity

vw m·s-1 – Wind velocity
V V – Voltage
Vd m3 – Droplet volume
VN - – Viscosity number
abbreviation description
CMR – Charge to Mass Ratio
CE – Charging Electrode
EHDA – Electrohydrodynamic Atomisation
ELPI – Electrical Low Pressure Impactor
EPI – EWICON Performance Index
EWICON – Electrostatic Wind Energy Converter
GSD – Geometric Standard Deviation
HPMS – High Pressure Monodisperse Spraying
HVDC – High Voltage Direct Current
LPDA – Laser Phase-Doppler Analyser
SE – Steering Electrode
164
Acknowledgements
List of references
[1] International Energy Agency, World Energy Outlook 2006, pp.66-67, 2006
[4] Federal Ministry for the Environment, Nature Conservation and Nuclear
Safety, “Environmental Policy – Nuclear Safety – Report under the Convention
on Nuclear Safety by the Government of the Federal Republic of Germany”,
Public Relations Division, pp.1-5, September 2001
[5] International Energy Agency, World Energy Outlook 2006, pp.265, 2006
[6] International Energy Agency, Renewables in Global Energy Supply, An IEA

Fact Sheet, 2007
[7] Renewable Energy Policy Network for the 21st Century (REN21), Renewables
2007 Global Status Report, Paris: REN21 Secretariat and Washington, DC:
Worldwatch Institute, Deutsche Gesellschaft für Technische Zusammenarbeit
(GTZ) GmbH, 2008
[8] Ministerie van VROM, brochure “Het kabinet zet in op windenergie”, juni
2008
[9] Koninklijk Instituut van Ingenieurs, “Windenergie is een politieke keuze:

Technisch haalbaar maar niet rendabel”, Persbericht referentienummer:
An963049.P20, 2002
[10] Williams, P., Lansdorp, B., and Ockels, W., “Flexible Tethered Kite with
Moveable Attachment Points, Part II: State and Wind Estimation,” AIAA
Atmospheric Flight Mechanics Conference, 2007
[11] Keith, D.W., DeCarolis, J.F., Denkenberger, D.C., Lenschow, D.H.,

Malyshev, S.L., Pascala, S., Rasch, P.J., “The influence of large-scale wind
power on global climate”, PNAS, volume 101, no. 46, 2006
[12] Marks, A.M., US Patent No. 4.206.396 and 4.433.248

165
[13] Visschedijk, B, Graduate thesis “Electrostatic Generators”, NanoStructured

Materials Group, Faculty of Applied Sciences, Delft University of
Technology, 2005
[14] Sato, M., Nakayama, Y., Pan, R., Ueta, K., Shimada, Y., “Study on EHD
Wind Power Generation Utilizing Charged Water Droplets due to Contact to
Electrode”, International Symposium on Electrohydrodynamics, pp.33-36,
Argentina, 2006
[15] Young, A., Young, M., Scott, B., “High Voltage”, AC/DC, Australia, 1975
[16] Betz, A., “Das Maximum der theoretisch möglichen Ausnützung des Windes
durch Windmotoren”, Zeitschrift für das gesamte Turbinenwesen, vol. 26,
pp.307-309, 1920
[17] Lanchester, F.W., “A contribution to the theory of propulsion and the screw
propeller”, Transactions of the Institution of Naval Architects, vol. 57, pp.98-
116, 1915
[18] Hinds, W.C., Aerosol Technology: Properties, Behavior and Measurement of

Airborne Particles, Second Edition, New York, 1999
[19] Cloupeau, M. “Research on Wind Conversion by Electrofluid Dynamic

Processes”, Arch. Mech., vol. 5, no. 6, pp.529-538, 1986
[20] Lawson, M.O., von Ohain, H., Wattendorf, F., “Performance potentialities of
direct energy conversion processes between electrostatic and fluid dynamic
energy”, Aviation Research Laboratory, vol. 178, 1961
[21] Fuchs, N.A., Sutugin, A.G., “Generation and use of monodisperse aerosols.”
in Davies, C.N., ed. "Aerosol Science", Academic Press, pp.1-30, New York,
1966
[22] Bodega, R., Sonneveld, P.J., Morshuis, P.H.F., van Rooij, R., van Eldijk,
L.G.M., ”Conversion of wind into electrical energy without the need for
moving parts”, 8th International Symposium on High Voltage Engineering,
Millpress, Rotterdam, 2003
[23] Peskin, R.L., Raco, R.J., “Ultrasonic Atomization of Liquids”, Journal of

Acoustical. Society of America, vol. 35, pp.1378, 1963
166
Acknowledgements
[24] Teunou, E., Poncelet, D., “Rotary disc atomisation for microencapsulation
applications—prediction of the particle trajectories”, Journal of Food
Engineering, vol. 71, pp.345-353, 2006
[25] Kreuger, F.H., “Industrial High Voltage”, Delft University Press, 1991
[26] Kazkaz, G., “Electric field and space charge of spherical electrode at high
voltage concentric with a spherical ground conductive target”, IEEE
Transactions of Industry Applications, vol. 34, no. 4, 1998
[27] Grace, J.M., Marijnissen, J.C.M., “A review of liquid atomization by

electrical means”, Journal of Aerosol Sciences, vol. 30, no. 7, pp.1005-1019
[28] Hartman, R.P., “Electrohydrodynamic atomization in the cone-jet mode.

From physical modelling to powder production”, PhD thesis, Delft
University of Technology, 1998
[29] Borra, J.-P., Hartman, R., Marijnissen, J., Scarlett, B., “Destabilisation of
sprays in the cone-jet mode by electrical discharges on the jet”, Journal of
Aerosol Sciences, Suppl. 1, pp.203-204
[30] Cloupeau, M., Prunet-Foch, B., “Electrohydrodynamic Spraying Functioning

Modes: A Critical Review”, Journal of Aerosol Sciences, vol. 25, no. 6,
pp.1021-1036, 1994
[31] Gañán-Calvo, A.M., Davilla, J., Barrero A., “Current and droplet size in the
electrospraying of liquids. Scaling laws”, Journal of Aerosol Sciences, vol.
28, pp.249-275
[32] Keskinen, J., Pietarinen, K. and Lehtimäki, “Electrical Low Pressure

Impactor”, Journal of Aerosol Sciences, vol. 23, pp.353-360, 1992
[33] Lubetkin, S.D., Akhtar, M., “The variation of surface tension and contact
angle under applied pressure of dissolved gases and the effects of these
changes on the rate of bubble nucleation”, Journal of Colloid and Interface
Science, volume 180, number 1, pp. 43-60, June 1996
[34] Lastow, O., Balachandran, W., “Numerical and experimental study of

electrohydrodynamic atomization (EHDA) of conducting of conducting
liquids in the cone jet mode”, International Symposium on
Electrohydrodynamics, pp.265-268, Argentina, 2006
167
[35] Crowley, J.M., “Simple expressions for Force and Capacitance for a
Conductive Sphere near a Conductive Wall”, Paper D1, Proceedings of the
Electrostatics Society of America 2008 Annual Meeting on Electrostatics,
2008
[36] Nilsson, J., Palm, L., “Charging current measurements and charging
synchronization in continuous jet”, Review of Scientific Instruments, vol. 72,
no. 2, pp.1574-1579, 2001
[37] Taylor, G., “Disintegration of water drops in an electric field”, Proceedings

of the Royal Society of London. Series A, Mathematical and Physical
Sciences, vol. 280, no. 1382, pp.383-397, 1964
[38] Djairam, D., Morshuis, P.H.F., Smit, J.J., “Optimising electrode design and
positioning for EHDA produced particles in a EWICON”, Conference on
Electrical Insulation and Dielectric Phenomena 2007, Canada, pp.478-481,
2007
[39] Djairam, D., de Wit, A.T., Morshuis, P.H.F., Smit, J.J., “Controlled and
efficient electrohydrodynamic spraying of water in an EWICON”,
Conference on Electrical Insulation and Dielectric Phenomena 2008, Canada,
pp.169-172, 2008
[40] Theodore, A.D., EWICON report “Continued prototype design and testing of
self-adjusting multi-nozzle electrospray system using porous technique”,
NanoStructured Materials Group, Faculty of Applied Sciences, Delft
University of Technology, 2006
[41] Djairam, D., Nijdam, W., Balendonck, J., Morshuis, P.H.F., Smit, J.J. Smit,
“Converting wind energy to electrical energy using charged droplets in an
electric field”, Proceedings Conference on Electric Insulation and Dielectric
Phenomena 2007, Canada, pp.474-477, 2007
[42] Wever, R., “The effects of electric fields on circadian rhythmicity in men”,
American Journal of Epidemiology, vol. 109, pp.273-284, 1970
[43] Repacholi, M.H., Greenebaum, B., “Interaction of static and extremely low
frequency electric and magnetic fields with living systems: health effects and
research needs”, Bioelectromagnetics, vol. 20, no. 3, pp.133-160, 1999
168
Acknowledgements
Acknowledgements
“It is by chance that we met, by choice that we became friends.”
Writing a PhD thesis is often considered a solitary activity, but this work
could not have been even remotely possible, were it not for the help and support of
numerous people. These people have contributed in different ways: some did by
helping me with experiments, by participating in brain storm sessions and, on more
than one occasion, by pointing me in the right direction. Others did by just being
there while joking around over the occasional cup of coffee or glass of beer and,
thus, keeping me sane. Between the two groups, there was plenty of overlap and to
all of them I want say a few little words…
First of all, I am greatly thankful to my promotor, professor Johan Smit, for

his trust and the opportunity he has given me to work on this great project. I value
the many discussions we have had about the chapters of my thesis and I truly feel
that they have made this thesis much better.
I’d like to thank my daily supervisor, dr.ir. Peter Morshuis, for looking
after me all these years. Peter, not only did you help you me with the EWICON by
thinking with me and asking me the right questions, you also made me feel at home
at the group.
Edward, I still remember you asking during my job interview while you
were looking through my M.Sc. thesis, “What have you done for science?” After
an extra four years, I’m sure I can give a better answer now. Sander, besides being
a fun person to be around with, thanks for giving me the opportunity to teach. For
me, that was definitely one of the most rewarding aspects of doing a PhD. Maria,
thanks for making my life so much easier and all the “gezelligheid”!
Simulations, reading papers and thinking can only get you so far and none
of my experimental results would have been possible without the invaluable
support from ing. Paul van Nes, Aad van der Graaf and Bertus Naagen. Paul, I
appreciate your support and your efforts to instil some form of discipline in me.
Aad, thanks for all your pieces of art, “Wonderen doen we direct, het onmogelijke
duurt iets langer!” Bertus, thanks for all the practical help during my experiments
and for occasionally saving my life when I was sprinting towards my set-up in
great enthusiasm.
169
The EWICON project group consisted of many people with whom I

enjoyed working together. Anna Hubacz, Wietze Nijdam, Jan Marijnissen, Jos
Balendonck, Piet Sonneveld, Alex Theodore and Ruud van Rooij. Together, we
made important steps!
To me, a working environment with good atmosphere is just as important

as the job itself and the High Voltage Group provided that in abundance. Riccardo,
thanks for helping me get started with the EWICON and for being my roommate
for 3.5 years. My current roommates, Thomas and Roman, I’m glad we all like
music! Roma, in de toekomst gaan we alleen Nederlands praten! Thomas, thanks
for being the fellow geek/bassist/movie buff of the group. Arie, the graduation
work you did for the EWICON was as impressive as it was valuable.
My former fellow PhD’s who have always been open for spirited
conversation, Ben, Tjerk, Robert, Belma and Harry. My current fellow PhD’s,
Tom, Gautam, Roy, Piotr, Muhannad and Qikai, keep the group alive and kicking!
Pantelis, we still have to do a reunion of the High Voltage Band and I hope
to see you graduate after me! Rogier, my fellow Dutch guy, together we made sure
that the farmer’s accent was regularly heard. I can’t thank you enough for all the
help with my research and thesis.
Guys, I’ve had a great time with you all, drinking beers in the /pub, all the
Bond nights, doing sports together and just hanging out in general!
I’ve been wandering around in Delft for almost 14 years now and, along
the way, I was lucky enough to meet a lot of people who I am fortunate enough to
call my friends. My first mention goes out to my first student house “de
Blokkendoos”, both the old and new members, where I am still always welcome to
unwind, even though I left the house six years ago. Vincent, Eva, Vincent, Arjen,
Sieb, Menno, Widy, Edwin, Eelco, Tomas, Kristian, Piet, Hans, Pieter, Robert,
Sjoerd, Bastiaan, Melanie, Rosemarijn. Dinners, parties, holidays… those were and
still are good times! Tomas, just like with my M.Sc. thesis, many many many
thanks for all the graphics works you did!
Martial arts will always be an influential part of my life and so will

Yoroshi and all the people I’ve met through that. Sensei Dick, thanks for almost ten
years of guidance. Linda, Hadewig, Dedmer, Boi, Gerard, Mohammed, Jeroen,
Michael, see you either on the tatami or the slopes!
Sebastiaan, Hanneke, Ronald, Saskia, here’s to many more dinners… and
maybe some jogging!
Mathijs, Alwin, Joël, movie night soon? I’ll do the cooking!
Esmee, Arvid, Nathan, my Donner buddies. It seems like ages ago that we
were thrown together in the bookstore and even though we worked for a brief time
and are all very different, we still can’t seem to shake each other!
170
Acknowledgements
Piet and Griet, we’ve know each other for over ten years now and during
that time, you’ve given me the opportunity to explore Leiden and this exotic
subculture of bearded archaeologists. I wish you all the best with Jasper!
Becky, thirteen years of flowery English, that’s the reason why my skills
progressed beyond high school levels.
Luke and Janneke, visiting New Zealand and seeing you guys get married
was definitely one of the best moments of my life. Is Dilan already trying to
outsmart you?
Anne, you will go wherever your saxophone will take you, but that has
never stopped us from seeing each other and trying to cram as much as possible in
the few moments we have.
Mascha, ever since you set foot on the Blokkendoos, we have been great
friends and shared many things and that never changed, not even when you moved
to Eindhoven! Good luck with the new house and of course, Philip’s PhD defence.
Bas, we’ve been friends since we were thirteen, so what can I say… we’re
both from the northern parts, we’re both men of few and many words, so when
we’ve got something to say, we talk and otherwise we play guitar. That level of
comfort is rare and priceless. Now, a new phase has started for you together with
Micheline and Midas. Thanks for taking me along for the ride!
Marnix, physics buddy, guitar buddy, partner-in-crime, we shared our high
points and low points in life, exams, graduation, break-ups, family matters, breath
taking holidays, band practice and much more. The only thing I can do is return the
compliment: you have been a wonderful friend! The fact that we will graduate at
the same point in our lives is just plain freaky and way cool!
The last part is reserved for the most important people in my life, my
family. My grandparents and my many aunts and uncles, cousins, nieces and
nephews have all contributed in their little way with advice and encouragements.
The knowledge, that I would always have a home to back to no matter

what would happen in life, has given me strength and courage to do whatever I
wanted.
Rishi, bedankt voor het zijn van een goed broertje, van een jong opstandig
pubertje tot een verstandig volwassen man. We lijken in niets op elkaar en toch sta
je altijd voor me klaar… zoals de vele autotrips. Sangieta, bedankt voor het zijn
een onwijs lief zusje, ook van een jong opstandig pubertje tot een verstandig
volwassen mooie vrouw. Zo vaak als je voor mij gezorgd heb met alles zal ik nooit
vergeten.
Lieve pa en ma, door jullie onvoorwaardelijke steun en liefde, door jullie
aanmoediging om the leren, al van heel jongs af aan, heb ik dit kunnen bereiken.
Erg bedankt!
171
Curriculum Vitae
Dhiradj Djairam was born on May 19th, 1977 in

Groningen, the Netherlands. He received his secondary
school degree from the Praedinius Gymnasium in
Groningen in 1995. He received his M.Sc. degree in
Applied Physics from the Delft University of Technology
in 2002, specialising in experimental and laser optics.
He started his Ph.D. research project in 2004, called

“EWICON II – Electrostatic Wind Energy Converter”, in
which he has investigated an alternative method of
extracting energy from the wind. He conducted this research, which led to this
thesis, at the High Voltage Technology and Management Group, Delft University
of Technology.
Currently, he works as a post-doc on the SINERGIE Smart Grids project, also at

the High Voltage Technology and Management group, Delft University of
Technology.
173

EWICON

Uploaded by

Document Informationclick to expand document informationWind Power static turbine in Delft, Netherlands. College research paper

Document Informationclick to expand document information

Copyright:

Available Formats

EWICON

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EWICON

Uploaded by

Copyright:

Available Formats

The Electrostatic Wind Energy

ter verkrijging van de graad van doctor

Rector Magnificus Voorzitter

Copyright © 2008 by D.Djairam

Cover: An artist’s impression of an EWICON system at sea.

Aan mijn ouders

In chapter 2, the theoretical framework that governs the operation of the

In chapter 3, a number of spraying methods to create charged particles are

In chapter 5, various versions of the EWICON system are tested while

Door de wereldwijde toenemende vraag naar energie is het onderzoek naar

In hoofdstuk 1 wordt de huidige mondiale energiesituatie uiteengezet

In hoofdstuk 2 wordt het theoretisch fundament van de werking van het

In hoofdstuk 3 wordt een aantal verschillende methoden behandeld

In hoofdstuk 4 wordt het ontwerp van het algehele EWICON-testsysteem

In hoofdstuk 5 worden verschillen versies van het EWICON-systeem

In hoofdstuk 6 wordt dit proefschrift afgesloten met de conclusie dat,

Summary ................................................................................................................ vii

2. Theoretical basis and modelling ......................................................................... 19

3. Charged droplet creation methods...................................................................... 43

3.1.4 Operable under windy condition .................................................. 47

4. EWICON system design..................................................................................... 75

6. Conclusions and recommendations .................................................................. 133

Appendix A. Betz’ law ................................................................................ 139

Appendix B. Evaporation and relaxation times of droplets......................... 143

Appendix C. Numerical modelling in MATLAB........................................ 147

Appendix D. Statistical analysis in MATLAB ............................................ 157

Appendix E. LORENTZ ............................................................................. 161

List of symbols and abbreviations ........................................................................ 163

List of references .................................................................................................. 165

Acknowledgments ................................................................................................ 169

Curriculum Vitae .................................................................................................. 173

While wind energy is often viewed as an energy source on its own,

1.1 Energy production in general

Worldwide energy consumption is growing and as countries like China and

World Primary Energy Demand

Of the current global energy production 86 percent is delivered using fossil

Both forms of energy production have their own drawbacks considered

Similar considerations can be made concerning fission based nuclear

1.2 Alternative energy sources

Considering the issues with the aforementioned sources of energy,

- hydrogen/ethanol fuel cells,

World Renewable Energy 2006

Geothermal power (0.97%)

Solar PV, grid-connected (0.52%)

Solar PV, off-grid (0.28%)

Ocean (tidal) power (0,03%)

1.3 Wind energy in general

Usually, when it comes to alternative means of generating energy, several

Sustainable energy: A form of energy generation which can be

Renewable energy: A form of energy generation in which the required

In the Netherlands, the installed wind power capacity currently is

1.3.1 Conventional methods

Another rotation based method is the Darrieus wind turbine, shown in

Figure 1.5. An example of a Darrieus wind turbine.

1.3.2 Drawbacks of wind turbines

1.3.3 Alternative wind energy methods

1.3.4 Use of wind energy

Also, computational research using meteorological models as described in