V Weeda MscThesis

Solar Powered Infotainment Spot
The design, feasibility study and fabrication of an au-

tonomous PV system
V. Weeda
Master of Science Thesis
Photovoltaic Materials and Devices

The design, feasibility study and fabrication of an autonomous PV
system
Master of Science Thesis
For the degree of Master of Science in Sustainable Energy Technology

at Delft University of Technology
V. Weeda
October 31, 2014
Faculty of Applied Sciences · Delft University of Technology

Copyright © Photovoltaic Materials and Devices (PVMD)
All rights reserved.
Delft University of Technology
Department of
Photovoltaic Materials and Devices (PVMD)
The undersigned hereby certify that they have read and recommend to the Faculty of
Applied Sciences for acceptance a thesis entitled
by
V. Weeda
in partial fulfillment of the requirements for the degree of
Master of Science Sustainable Energy Technology
Dated: October 31, 2014
Supervisor(s):
Dr. O. Isabella
Prof. Dr. M. Zeman
Reader(s):
Dr. L. M. Ramírez Elizondo
Abstract
Information technology has recently eased the access to all kinds of information. For example,
the PVMD group at TU Delft, via its website, shares information on its research and educa-
tion. However, one could argue that not everybody in the university campus knows the latest
developments in the PV industry or the activities of the PVMD group. This has driven the
desire for an infotainment spot (information + entertainment) in the campus of the university.
The demonstrator provides information to people on campus via a rugged touchscreen that
is powered by solar energy. In such an autonomous system characterized by a minimal loss
of load probability, a flexible CIGS module is deployed as market alternative to rigid c-Si
modules. Data from a weather station of KNMI was used to accurately estimate the amount
of solar energy available in the Netherlands. The influence of the surroundings of the info-
tainment spot on the amount of all irradiance components was then accordingly investigated.
Furthermore, a methodology was developed to make an accurate estimate of the solar irra-
diance on a curved surface, since the key component in the design of the infotainment spot
was the flexibility of the PV panel. Knowing the irradiance on the PV module, its power
output was estimated by a PV model taking temperature, wind and irradiance effects into
account. Consequently, the system topology was designed to have the smallest possible losses
and highest yield of PV power. A load profile was developed in combination with measure-
ments of actual power consumption to simulate the system performance throughout the year.
The study indicated that the meteorological data guarantees an average annual minimal
horizontal irradiation of 1000 kWh m-2 in the Netherlands. An extensive analysis of the
surroundings was performed with a 3-D model resulting in a sky view factor for each part
of the PV module and a shading coefficient as function of time. With the temperature and
irradiance effects taken into account, the energy production of the 90-W rated CIGS module
is 62 kWh/year, compared to an energy consumption by the loads of 19.3 kWh/year. The
12-V DC system is completely autonomous with a battery of 120 Ah, regulated by a charge
controller. The system can be adjusted in terms of rated power of the PV module and the
battery size if the infotainment spot were to be installed in a different location.
Master of Science Thesis V. Weeda

ii
V. Weeda Master of Science Thesis

Contents
1 Introduction 1
1-1 Energy Crisis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1-2 Information Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1-3 Motivation of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1-3-1 Infotainment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1-3-2 Scope of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1-4 Research questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1-5 Thesis outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Design & Location 7

2-1 Embodiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2-1-1 Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2-1-2 Touchscreen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2-1-3 Screen position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2-2 Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2-3 Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2-3-1 Irradiance measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2-3-2 Proposed locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3 Energy resources 19
3-1 Solar radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3-2 Meteorological data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3-3 Irradiance on a horizontal plane . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3-3-1 Irradiation on a horizontal plane . . . . . . . . . . . . . . . . . . . . . . 22
3-4 Irradiance on a tilted plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3-4-1 Direct irradiance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3-4-2 Diffuse irradiance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3-4-3 Albedo irradiance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3-4-4 Shading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3-4-5 Summation of the components . . . . . . . . . . . . . . . . . . . . . . . 37
3-5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

iv Contents
4 PV Potential 43
4-1 Curvature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4-1-1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4-1-2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4-1-3 Available energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4-2 PV output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4-2-1 The photovoltaic effect . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4-2-2 Irradiance dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4-2-3 Temperature dependence . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4-3 PV module output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4-4 Flexible PV modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4-4-1 Powerfoil® 115 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4-4-2 Powerfoil® 165 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4-4-3 Powerflex™ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4-4-4 DC yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4-5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5 Building the system 65

5-1 Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5-2 Load profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5-2-1 Power demand Toughpad FZ-A1 . . . . . . . . . . . . . . . . . . . . . . 67
5-2-2 Power demand USB ports . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5-3 System modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5-3-1 Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5-3-2 Charge controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5-3-3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5-3-4 System performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5-4 Building the infotainment spot . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5-4-1 The prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6 Conclusions and Recommendations 87

6-1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6-2 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6-2-1 Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6-2-2 Design & Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6-2-3 Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6-2-4 Additional loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6-2-5 Battery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
A Appendices 93
Bibliography 101
Glossary 105
List of Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
List of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

List of Figures
1-1 Annual PV electricity production. . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1-2 Information kiosk on a train station in Perth, Australia. . . . . . . . . . . . . . . 3
1-3 The location: TU Delft campus. . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1-4 MIT ‘soft rocker’: a solar charging station combined with a rocking chair. . . . . 5
2-1 The two concepts for the embodiment design of the infotainment spot. . . . . . 9
2-2 The Panasonic Toughpad FZ-A1. . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2-3 The tilt angle of the touchscreen. . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2-4 The orientation of existing touchscreens in typical Dutch public places. . . . . . 13
2-5 The design of the web page. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2-6 The vicinity of the EEMCS building. . . . . . . . . . . . . . . . . . . . . . . . . 15
2-7 Irradiation on the vicinity of the faculty building. . . . . . . . . . . . . . . . . . 16
2-8 Students spending their lunch break on the TU Delft campus. . . . . . . . . . . 17
2-9 The two proposed locations for the infotainment spot. . . . . . . . . . . . . . . 18
2-10 The two proposed locations seen from above. . . . . . . . . . . . . . . . . . . . 18
3-1 The components of irradiance on a tilted surface. . . . . . . . . . . . . . . . . . 20

3-2 Measuring solar irradiance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3-3 Global Horizontal Irradiance in the years 2012 and 2013. . . . . . . . . . . . . . 22
3-4 Irradiance levels on a sunny and a cloudy day. . . . . . . . . . . . . . . . . . . . 23
3-5 Definition of azimuth, altitude and tilt angle. . . . . . . . . . . . . . . . . . . . 25
3-6 The sunpath for different days in the year. . . . . . . . . . . . . . . . . . . . . . 26
3-7 Direct irradiation on a tilted plane. . . . . . . . . . . . . . . . . . . . . . . . . . 27
3-8 The influence of tilt angle on the SVF. . . . . . . . . . . . . . . . . . . . . . . . 27
3-9 Diffuse irradiation on a tilted plane. . . . . . . . . . . . . . . . . . . . . . . . . 29
3-10 Sky View Factor as function of tilt angle. . . . . . . . . . . . . . . . . . . . . . 30

vi List of Figures
3-11 Diffuse irradiation on a tilted plane, location-specific SV F . . . . . . . . . . . . 31

3-12 Setup for albedo measurements on campus grounds. . . . . . . . . . . . . . . . . 32
3-13 Albedo irradiation on a tilted plane. . . . . . . . . . . . . . . . . . . . . . . . . 33
3-14 The infotainment spot modeled in Google SketchUp. . . . . . . . . . . . . . . . 34
3-15 Creating the shading coefficient on August 12th 2014 using SketchUp. . . . . . . 35
3-16 Global irradiance on a tilted plane for two different shading models. . . . . . . . 37
3-17 Cumulative global irradiation on a tilted plane for two different shading models. . 38
3-18 Global irradiation on a tilted plane without surroundings. . . . . . . . . . . . . . 39
3-19 Global irradiation on a tilted plane on location 1. . . . . . . . . . . . . . . . . . 40
3-20 Diffuse irradiation + Albedo irradiation for different SV F s. . . . . . . . . . . . . 41
4-1 Definition of variables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4-2 Division of the curved plane into flat segments. . . . . . . . . . . . . . . . . . . 46
4-3 The top and bottom angle of segment i. . . . . . . . . . . . . . . . . . . . . . . 47
4-4 The marginal yearly irradiation on the surface. . . . . . . . . . . . . . . . . . . . 49
4-5 The deviation in Eaverage,global
0 compared with 104 segments as function of n . . 49
4-6 The deviation in Eaverage,global
0 compared with 104 segments as function of F . . 50
4-7 The offset margins on top of and below the module. . . . . . . . . . . . . . . . 51
4-8 Eaverage,global as function of F and Amodule . . . . . . . . . . . . . . . . . . . . 51
4-9 Irradiance level on the different segments in different times of year. . . . . . . . . 53
4-10 Schematic PV module configurations. . . . . . . . . . . . . . . . . . . . . . . . 53
4-11 The effect of irradiance and temperature on the IV-curve of a solar cell. . . . . . 55
4-12 Tambient throughout the year 2013. . . . . . . . . . . . . . . . . . . . . . . . . . 56
4-13 Wind speed throughout the first week of November 2013. . . . . . . . . . . . . . 57
4-14 Tcell as function of wind speed, for Tambient = 20o C and Gcell = 800W m−2 . . . 59
4-15 Tmodule during September 5th 2013. . . . . . . . . . . . . . . . . . . . . . . . . 59
4-16 Emodule as function of footprint F . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5-1 Topology of the system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5-2 Setup to measure power demand of the toughpad FZ-A1. . . . . . . . . . . . . . 67
5-3 Power demand of the toughpad FZ-A1 in automatic brightness mode. . . . . . . 68
5-4 Load profile of the tablet on a ’Summer Weekday’. . . . . . . . . . . . . . . . . 69
5-5 Setup to measure power demand of the USB ports. . . . . . . . . . . . . . . . . 69
5-6 Power consumption of the USB charger as function of the battery SOC. . . . . . 70
5-7 Load profile of the USB charger on a ’Summer Weekday’. . . . . . . . . . . . . . 71
5-8 Total daily load profile for a ’Summer Weekday’. . . . . . . . . . . . . . . . . . . 72
5-9 The effect of temperature on battery capacity. . . . . . . . . . . . . . . . . . . . 74
5-10 Setup to measure efficiency of the charge controller. . . . . . . . . . . . . . . . . 74
5-11 Charge controller efficiency as function of the load power. . . . . . . . . . . . . . 75

List of Figures vii
5-12 Definition of the energy flows in/from the charge controller. . . . . . . . . . . . 76

5-13 Control strategy for the charge controller. . . . . . . . . . . . . . . . . . . . . . 76
5-14 Typical variations in SOC of the battery throughout the year. . . . . . . . . . . . 78
5-15 The SOC in January for an 80 Ah battery. . . . . . . . . . . . . . . . . . . . . . 80
5-16 The influence of DOD and temperature on battery cycle life. . . . . . . . . . . . 81
5-17 Battery size and capacity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5-18 The components used in the prototype. . . . . . . . . . . . . . . . . . . . . . . 83
5-19 The skeleton construction of the prototype. . . . . . . . . . . . . . . . . . . . . 85
5-20 The prototype of the infotainment spot. . . . . . . . . . . . . . . . . . . . . . . 86

viii List of Figures

List of Tables
2-1 Physical dimensions of Dutch adults. . . . . . . . . . . . . . . . . . . . . . . . . 12
3-1 Irradiation on a horizontal surface in 2012 and 2013. . . . . . . . . . . . . . . . 23

3-2 Typical albedo values for various ground types. . . . . . . . . . . . . . . . . . . 32
3-3 Albedo on TU Delft campus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3-4 Yearly irradiation predicted with two shading models. . . . . . . . . . . . . . . . 37
3-5 Irradiation in 2013. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4-1 DC yield prediction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5-1 DC voltage range of the tablet and a typical 12V charge controller. . . . . . . . 66
5-2 Power consumption of the toughpad FZ-A1. . . . . . . . . . . . . . . . . . . . . 68
5-3 Power consumption of the USB charger. . . . . . . . . . . . . . . . . . . . . . . 70
5-4 Assumed usage of the infotainment spot. . . . . . . . . . . . . . . . . . . . . . . 71
5-5 Annual energy demand by the loads. . . . . . . . . . . . . . . . . . . . . . . . . 72
5-6 Estimate of battery size depending on screen brightness setting. . . . . . . . . . 77
5-7 Estimate of battery size depending on the PV module. . . . . . . . . . . . . . . 78
5-8 System performance as a function of battery size. . . . . . . . . . . . . . . . . . 79
5-9 Indication of the cost-price of the system components. . . . . . . . . . . . . . . 84

x List of Tables

Chapter 1
Introduction
1-1 Energy Crisis
Mankind has been mainly dependent on fossil fuels (oil, gas and coal) for energy production
over the last century. The availability of these fuels is however limited and varies greatly with
geographical location. This has made certain nations dependent on external energy sources
and therefore vulnerable to energy supply, fuel price and currency fluctuations. Furthermore,
the use of fossil fuels results in increasing levels of carbon-dioxide (CO2 ) and other greenhouse
gases (GHGs) in the atmosphere, causing the increase in greenhouse effect (Global Warming).
For these reasons, nations are more and more trying to shift to renewable energy sources
for their energy supply. One of the fastest growing energy technologies is photovoltaic (PV)
energy conversion, i.e. the direct conversion of sunlight into electricity. The global PV market
has experienced vibrant growth for more than a decade with an average annual growth rate
of 40%. The cumulative installed PV power capacity has grown from 0.1 GW in 1992 to over
100 GW in 2013, as shown in Figure 1-1. The International Energy Agency (IEA) expects
that 11% of the total electricity generation will be provided by PV by the year 2050 [1], while
nowadays this is only 0.1% [2].
The PV technology is relatively young with its first real market applications starting in the
1950s. In the last decades a lot of research has been done to improve the technology, which
has led to the wide acceptance and implementation of PV cells in the market. Delft University
of Technology (TU Delft) is one of the institutes performing research for further development
of the PV technology within the research group Photovoltaic Materials and Devices (PVMD).
1-2 Information Age
Aside from the developments in the PV technology in the last decades, a much bigger change
has happened: as Humbert [4] described, in the 1970s the developed world entered into a new

2 Introduction
Figure 1-1: Annual PV electricity production, adapted from: [3].
age, one of information. Since then, the Information and Communications Technology (ICT)
has totally changed our way of life, leisure and our means of communication and informa-
tion. Internet is obviously the best example of this technological revolution. Nowadays it is
very common that people in developed countries have access to internet and thus to incred-
ible amounts of information. Furthermore the access to internet has been made a lot easier
through the introduction of wireless networks and the use of smartphones. In the Netherlands
73% of the mobile phone users in the end of 2013 owned a smartphone [5], with which any
desired information on the internet can be accessed anywhere, anytime. The developments
in ICT have thus contributed to the fact that people nowadays are used to have access to all
this information.
This development has also changed the way that companies and organizations are providing
their target audience with information. To cut down on costs, manned service desks have
often made place for information kiosks, for example on train stations as shown in Figure 1-2.
1-3 Motivation of the thesis
To share information on its research and education, the PVMD group of TU Delft has pro-
vided detailed information on the designated part of the TU Delft website [7]. However, one
could argue that on the TU Delft campus not everybody knows what the latest developments
in the PV industry are, what the activities of the PVMD group contain or that this research
group of TU Delft even exists.
One of the main motivations for this thesis project is therefore to increase this “PV awareness”
on the campus of TU Delft. This thesis project contains the design and feasibility study of
an infotainment spot (see section 1-3-1) a demonstrator project with two main purposes:

1-3 Motivation of the thesis 3
Figure 1-2: Information kiosk on a train station in Perth, Australia [6].
Figure 1-3: The location: TU Delft campus [8].

4 Introduction
1. Demonstrating what is currently possible in the PV technology with the implementation

of a flexible PV module.
2. Providing the people on the TU Delft campus with relevant information and entertain-
ment via a touchscreen.
1-3-1 Infotainment
Previously the term infotainment is used, which is a combination of the words information
and entertainment. The goal of the infotainment spot is to provide both information and
entertainment to students, employees and visitors of TU Delft. The provided information
will be of three kinds:
• Practical information on TU Delft and the faculty Electrical Engineering, Mathematics

and Computer Science (EEMCS). A student for instance gets the opportunity to see
on campus where in the EEMCS building a computer is available, or check the latest
news on the TU Delft website.
• Information about solar energy and in particular the activities of the PVMD research
group. One can acquire information on the educational program on Solar Energy that
is offered, the research projects, but also simulate the benefits of a PV system on their
own house.
• The latest news, weather and traffic information. A visitor of TU Delft can easily find
out what time the tram from campus to Delft train station is leaving, or see weather
the traffic is jammed on the way home.
For the entertainment part one could think of a speaker that can be connected via bluetooth
to a user’s smartphone or tablet. Another possibility is to include a WiFi hotspot, so the
user can freely connect to the internet on the campus. A variety of games can be included
as well; not only on the touchscreen itself, but one could think of a social game on campus,
where the infotainment spot could function as game-master.
1-3-2 Scope of the thesis
This thesis project focuses on the design and feasibility of the infotainment spot powered by
solar energy. The ‘information’ section that will be provided by the touchscreen is included in
this project, whereas the ‘entertainment’ part is not covered and left for further development.
Another functionality of the infotainment spot is however included in this thesis, which is
the possibility for the user to charge his/her phone or tablet for free with purely solar energy
via a USB port. This functionality is already implemented in quite some places, e.g. on the
campus of Vanderbilt University in Nashville [9] and MIT in Massachusetts [10].
This project focuses merely on the feasibility study and will be concluded with a prototype
of the infotainment spot. Further optimization and finally the realization is outside the scope
of this thesis.

1-4 Research questions 5
Figure 1-4: MIT ‘soft rocker’: a solar charging station combined with a rocking chair [10].
1-4 Research questions
In the process of the design and feasibility study of an infotainment spot, certain problems
need to be solved and certain questions need to be answered. The following questions form
the core of the research that is performed in the framework of this thesis:
1. How much of solar energy can be harnessed by an existing flexible PV module based in
the Netherlands?
(a) How much of solar energy is available throughout the year in the Netherlands?
(b) In what way can the amount of solar energy incident on a curved plane be accu-
rately estimated?
(c) How accurate is the estimation of diffuse irradiance with the general models found
in literature?
(d) How much of electrical power can be generated by a PV module throughout the
year?
2. Is it possible to rely entirely on the solar power throughout the year to meet the daily
load demand?
(a) How does an autonomous PV system fare in terms of meeting the energy demand
of the touchscreen of the infotainment spot throughout the year?
(b) To what extent is it possible to include additional electronic loads apart from the
touchscreen?

6 Introduction
1-5 Thesis outline
This thesis is divided into a number of chapters, each describing a specific part of the project.
Below, an overview of the different chapters in this thesis given.
Chapter 1 - Introduction
This chapter provides an overview of the motivation and defines the research questions.
Chapter 2 - Design & Location

This chapter explains what design considerations have been taken into account for the embod-
iment of the infotainment spot. This chapter also describes the process of finding a desired
location for the infotainment spot.
Chapter 3 - Energy Resources

This chapter aims at accurately determining the amount of solar energy annually available
for use in the Netherlands.
Chapter 4 - PV Potential
A methodology is discussed to accurately determine the solar energy incident on a curved
plane. The PV model described in this chapter also takes into account the irradiance and
temperature effects on the output of a PV module.
Chapter 5 - System modeling

This chapter analyzes the reliability of the infotainment spot throughout the year for certain
load profiles. Extensive modeling is carried out and the storage behavior is analyzed under
varying conditions. It describes all the components that the infotainment spot consists of.
Chapter 6 - Conclusions and Recommendations

This chapter summarizes the key results of this project and answers the research questions.
Furthermore, it lists some recommendations for further development of the infotainment spot.

Chapter 2
Design & Location
The infotainment spot is meant to be a demonstrator project on the campus of Delft University
of Technology (TU Delft), in the vicinity of the faculty of Electrical Engineering, Mathematics
and Computer Science (EEMCS). Both the design and the location of the infotainment spot
has to be thought through, in order for people to make the most use of it. In the following
sections the consideration for both the design and the location are discussed.
2-1 Embodiment
One of the key aspects of a design project is the list of demands and wishes. The demands
represent the features that ought to be implemented in the design for sure, whereas the wishes
are seen as a nice bonus in the design, but not a necessity. The list of demands and wishes
for the infotainment spot is described below.
Demands
• Touchscreen
The information (and entertainment) is provided via a touchscreen.
• PV Module
The touchscreen is powered by a flexible photovoltaic (PV) module.
• Vandalism
The infotainment spot is vandalism-proof and theft-proof.
• Weather
The infotainment spot can withstand all realistic weather conditions.
• Design
The design of the infotainment spot has to be attractive for users. Furthermore, the
flexibility of the PV module should be demonstrated with the design.

8 Design & Location
• Surroundings
The infotainment spot should fit well in the surroundings, i.e. on the campus of
TU Delft.
Wishes
• USB port
The infotainment spot offers the user an option to charge a phone or tablet via a USB
port.
• Other loads
The infotainment spot has other electronic loads, such as a LED light strip or a bluetooth
speaker.
• WiFi
The infotainment spot functions as a WiFi hotspot to which users can connect freely
with their smartphone, tablet or laptop.
2-1-1 Concepts
One of the biggest challenges is to make an elegant, attractive concept, which is vandalism-
and theft-proof, as usually the latter results in plain, unattractive products. Sketches of differ-
ent ideas on the design have been made and two concepts were chosen as realistic possibilities
for the final design. The two concepts are shown in Figure 2-1. At this stage the concepts
are not very detailed yet, only one of them is chosen to be further developed.
The umbrella (Figure 2-1a) consists of a large vertical pole to which the touchscreen is at-
tached. On top of the pole, an umbrella-kind of marquise is giving shelter for the user in
case of rain, as well as a place to mount the flexible PV module. A place to store the electric
components such as batteries is created at the bottom of the pole, at the same time providing
more stability to the construction.
The lectern (Figure 2-1b) is a more compact design with no components above the touch-
screen. Instead, the flexible PV module is mounted on the backside of the construction,
providing a better view at the module for people that are interested. The body in which the
touchscreen is embedded provides space for storage of the electric components.
Comparison
The two concepts have implemented all the demands and all wishes can be added to both of
them. In terms of the vandalism and theft demand, the umbrella scores better, as the PV
module cannot be accessed by a user without serious efforts. The marquise-feature however
comes with a challenge too, as stability of the construction is very important, especially in
the vicinity of the EEMCS building that is known for its windy conditions.

2-1 Embodiment 9
(a) Concept A: The umbrella (b) Concept B: The lectern
Figure 2-1: The two concepts for the embodiment design of the infotainment spot.
The lectern obviously gives easier access to the PV module, so in this aspect some additional
safety measures have to be taken to prevent the PV module from being stolen. However,
the rigid structure of the concept makes it favorable in terms of stability. Furthermore,
the flexibility of the PV module is clearly visible and the concept is more elegant than the
umbrella due to its dimensions; the umbrella should be at least 2.5m high to provide enough
space for users to access the touchscreen. The smaller type of design fits therefore better
in the surroundings. For these reasons the lectern is chosen to be further developed as final
design of the infotainment spot.
2-1-2 Touchscreen
The information that will be available is provided to the user via a touchscreen. The choice
for a certain type of touchscreen is based upon its demands and wishes.
Demands
• Operation
The touchscreen can be operated by touching it with ones fingers, no other accessory is
needed.
• Weather
The touchscreen needs to be able to withstand rain, dust, high and low temperatures,
as it will be placed outdoors permanently. From now on, this will be referred to as a
rugged screen.

10 Design & Location
• Size
The touchscreen needs to have a screen diagonal of minimum 10 inches, so that a certain
website is easily displayed without the need for zooming in/out.
• Brightness
The touchscreen needs to have a high brightness level, as the surrounding brightness
outdoors can hinder the view of the screen drastically.
Wishes
• Variable brightness
When the brightness of the screen can be varied, either automatically or manually, this
could be an energy saving option.
• Power consumption
The power that the screen demands is as low as possible. This is of course preferable,
but possibly in contrast with the demands for the size and brightness of the screen.
Three general groups of rugged touchscreens are available on the market: rugged touch panel
computers, rugged touch monitor displays and rugged tablets. The screen size of the former
two groups is generally larger that those of the rugged tablets; up to 20” and possibly even
larger. This comes however with the biggest drawback of the rugged touch panel computers
and monitor displays, which is the power consumption: typical 15” and 21.5” rugged touch
monitor displays consume 32 W and 51 W respectively, the touch panel computers of the
same size even more [11]. In contrast, the typical power consumption of a rugged tablet
working in the same brightness level is less than 10 W [12] [13]. The power consumption of
the screen is important as it determines mainly the sizing of the other system components,
such as the PV module and the energy storage. The lower the power consumption of the
touchscreen, the smaller are the sizes and costs of the later components. This is the reason
that only rugged tablets are considered an option to be implemented in the infotainment spot.
Even though the market for rugged tablets isn’t as big as the one for non-rugged tablets still
some variety between tablets can be found. The rugged tablet market leader is Panasonic,
especially in the screen sizes of 10” and larger. All rugged tablets on the market with this
minimum size demand have been compared in terms of brightness, screen size, power con-
sumption and price. The larger 20 inch tablets, such as the UT-MB5 and UT-MA6, come
with the drawback of a high power consumption (approximately 15-20 W already at a very
low brightness setting of 60 nits). The 10 inch FZ-A1 and FZ-G1 tablets are comparable in
most options, with the later having more advanced options, resulting in a higher cost price
of ± €2750 [14] in comparison with ± €1000 for the FZ-A1 [15].
The choice for rugged touchscreen has therefore been the Panasonic Toughpad FZ-A1, see
Figure 2-2, an android based rugged tablet that meets all demands and wishes. The 10.1 inch
display has an anti-reflection and anti-glare treatment and is suitable for outside usage with
a high brightness of 500 nits. It has an IP65 sealed all weather design, meaning that it is
completely dust tight and can withstand 12.5 L min−1 of water jets without harmful effects.

2-1 Embodiment 11
Figure 2-2: The Panasonic Toughpad FZ-A1 [16].
Power management options are included in the form of sleep mode and automatic variable
brightness. The general power consumption is rated at 3.5 W; however, the brightness setting
(60 nits) in this situation is much lower than will be needed for the infotainment spot.
2-1-3 Screen position
The user of the infotainment spot has to touch the screen in order to get the desired infor-
mation displayed on the screen. To make the user experience of the touchscreen as good as
possible, the touchscreen should be positioned such that it is comfortable to operate; at the
same the information on the screen should be clearly visible for the user. The two parameters
to set, are the height of the screen and its tilt angle. The screen is comfortable to operate
when the height is approximately the elbow height of the user. However, for the information
on the screen to be clearly visible to the user, the screen should be higher than this, closer
to the eye-level of the user. Therefore a compromise should be made, resulting in a screen
height between the elbow- and eye-level of the user.
To determine the elbow- and eye-level of an average user, the anthropometric database Dined
has been consulted with Dutch adults (both male and female) from 20-60 years old as popula-
tion. According to the database both parameters are normally distributed. The results from
the Dined database, presented in Table 2-1, imply a restriction on the screen height between
1.08 m and 1.63 m.

Mean [mm] Standard deviation [mm]

Elbow height - standing 1084 74
Eye height - standing 1634 102
Table 2-1: Physical dimensions of Dutch adults [17].
Figure 2-3: The tilt angle of the touchscreen.
Another restriction on the screen orientation can be determined, namely on the screen tilt
angle (θscreen ). One of the design criteria is that sunlight cannot illuminate the touchscreen
directly, as this would make it unnecessarily hard for the user to read the information on the
screen.
As it is likely that the PV module will have South orientation to harvest as much solar energy
as possible, the touchscreen will face North. This means that the tilt angle of the screen
should be greater than the maximum altitude angle of the sun (asun ) - the angle that the
sun makes with respect to the horizon - as shown in Figure 2-3. Calculated with a Sun Path
Algorithm (SPA), explained in detail in 3-4-1, the maximum altitude of the sun in Delft is
61.4o . This implies a minimum screen tilt angle of 62o in order to prevent direct sunlight to
illuminate the screen.
Apart from setting restrictions on the screen height and tilt angle, the final design will have
both parameters fixed. In order to make a deliberate decision on both parameters, existing
touchscreens in the city Delft are investigated. From each of the 12 existing touchscreens
both parameters are measured; in Figure 2-4 the height of middle of the screen is plotted
versus the screen tilt angle. The black line for each screen represents the actual tilt angle.
The green box area shows an area where the screen tilt angle is smaller than the minimum
value of 62o .

2-2 Content 13
Figure 2-4: The orientation of existing touchscreens in typical Dutch public places.
From own experience it is determined that a small screen tilt angle gives higher comfort, as
the passe-partout - the edge enfolding the actual touchscreen - then gives a more comfortable
support while operating the screen. Therefore the screen tilt angle is determined to be 65o ,
close to the minimum of 62o .
The screen height is determined next, with the reference of existing screens and the restrictions
by Dined, resulting in 1.25 m. The final position of the screen is drawn in Figure 2-4 in cyan.
2-2 Content
As discussed in section 1-3-1 the informative content will be of three kinds: TU Delft and
EEMCS related information, information about solar energy (projects) and practical infor-
mation such as the latest news and the weather forecast. The information should be updated
automatically, which requires an internet connection. A very functional and straight forward
way of presenting the information is by showing the user a homepage with links to the des-
ignated informative web pages.
The homepage should be well-organized to provide the user a clear and quick view of all the
web pages available. The design of the homepage therefore is built out of buttons (or tiles),
each linked to a specific web page. The buttons are organized in three rows, each represent-
ing one of the classes of information. Each button consists of an icon - representative for the
linked web page - with a title below.
The purpose of this homepage is solely for the use in the infotainment spot. Therefore the
page is designed to precisely match the dimensions of the screen of the toughpad FZ-A1: 1024
x 768 pixels. The design of the page is shown in Figure 2-5. The homepage is written in html
and php and can be easily adjusted or updated remotely.

Figure 2-5: The design of the web page.

2-3 Location 15
(a) The front side of the building. (b) The back entrance.
Figure 2-6: The vicinity of the EEMCS building.
When this homepage is shown in a regular android web browser, the user can access not only
the designated pages, but also browse to other web pages - where the infotainment spot is
not meant for - and even access the settings of the tablet. To prevent this from happening,
the homepage will be shown in a so-called kiosk browser, an android application that makes
the functionalities of the user very limited: only the homepage is accessible (with the links to
the designated pages) and the browsers ‘home’, ‘forward’ and ‘back’ buttons. The settings of
the tablet can still be accessed, but is now password protected.
2-3 Location
The infotainment spot is a demonstrator project of the PVMD group of TU Delft, which
is part of the EEMCS faculty. Therefore its desired location is in the vicinity of the main
entrance of this faculty. The location of the main entrance however possibly poses a problem
for the working of the system, as this area suffers from shading for a large part of the day. In
the morning of a typical day the lower building of the faculty shades a part of the area, but
the worst shading is caused by the high building of the faculty in the afternoon. In principle,
a location with less shading would be at the backside of the high building, where there is also
an entrance.
2-3-1 Irradiance measurement
To get an idea on how the shading develops throughout the day, measurements in the vicinity
of the EEMCS faculty have been performed. On December 20th 2013, one of the ‘shortest’
days of the year, the solar irradiance was measured on several spots using a Hukseflux LP02
pyranometer at each hour from sunrise to sunset. At times that the buildings were shading a
certain spot, much lower irradiance was measured. Throughout the measurements there was
a clear sky, so shading on a location was purely caused by its surroundings.

Figure 2-7: Irradiation on the vicinity of the faculty building.
The results of the measurements can give a very rough approximation of the irradiation on a
typical winter day with clear sky, if the hourly measured irradiance is assumed to be constant
for that hour. The results of this first order estimate are shown in Figure 2-7.
It can be clearly seen that the spots close to the back entrance of the building are favorable
in terms of irradiation; only in the morning some shading was caused by the high building.
However, for the infotainment spot to be a good demonstrator project, it should be on a
location where more people will pass by so it can be optimally used. This means that even
though the irradiation is less than on the back side, a location on the front side of the high
building will be chosen.
The locations on the front side, close to the main entrance of the faculty have very low daily
irradiation, even though the measurements were performed under clear sky conditions. This
makes it very unsuitable for a solar powered system to be installed there. That is why a
location further away from the building is preferred, such as the one spot located close to the
tram stop.
The campus grounds are property of TU Delft and for the infotainment spot to be installed,
the Facility Management & Real Estate (FMRE) needs to be consulted. A proposal was
made for the infotainment spot and FMRE was willing to discuss the proposal further, as
long as the location would be on the grass of the campus grounds. To prevent users from the
need of walking through the grass to use the infotainment spot, a location next to an existing
footpath through the grass is the solution.

2-3 Location 17
Figure 2-8: Students spending their lunch break on the TU Delft campus.
In case of good weather this will not make much of a difference, as people on campus use the
grass itself as well as shown in Figure 2-8, but on rainy days, when the grass is wet or the
soil is muddy, people will less likely walk on the grass.
2-3-2 Proposed locations
Finally two proposals for a location are decided and presented in Figure 2-9 and Figure 2-10.
Both of them are in the grass of the campus grounds, next to a footpath. Both of the pathways
are leading to (or away from) the main entrance of EEMCS. In the determination of these
specific locations also the possible shading by the surrounding objects is thought of. The way
that shading influences the irradiation on the locations is discussed in detail in section 3-4-4.
Connectivity
The infotainment spot needs a reliable internet connection to function properly. The two
locations however, are located too far away from the EEMCS building to pick up a reliable
network (eduroam) signal. Both locations are actually located closer to the faculty of Civil
Engineering and Geosciences (CEG), but not close enough for a reliable eduroam signal from
that building either. If the infotainment is to be realized on one of the proposed locations,
then a WiFi amplifier needs to be installed in either one of the faculties (preferably CEG) to
guarantee proper functioning of the system.

(a) Proposed location 1. (b) Proposed location 2.
Figure 2-9: The two proposed locations for the infotainment spot.
Figure 2-10: The two proposed locations seen from above, adapted from [18].

Chapter 3
Energy resources
In the process of choosing a suitable location, the possible shading by surrounding objects
played a role. For the infotainment spot to work properly and continuously it is a necessity
that there is enough solar energy available to power the system. In this chapter a methodology
is discussed with which the available solar energy on both locations is accurately determined.
3-1 Solar radiation
Part of the radiation that the sun emits reaches the atmosphere of the earth. As the sunlight
travels through the earth’s atmosphere, it interacts with various particles that it finds on its
way towards earth’s surface. Some molecules will absorb the sunlight and thereby alter its
molecular vibrations and/or rotations. Other molecules will more likely scatter the incoming
radiation, changing its original path. It is for this reason that during a day with a clear sky
we actually can see the atmosphere and not the black sky that we see at night. One specific
form of scattering (Rayleigh scattering) in the atmosphere is much stronger for blue light
than for other colors, so that blue light gets scattered away from the original path of the sun
the strongest and we observe a blue sky.
Not only the sunlight that did not deviate from its original path will eventually fall onto
earth’s surface - referred to as direct radiation - but also this scattered light, which is called
diffuse radiation, will reach the earth’s surface. If the relevant surface is tilted and/or sur-
rounded by other objects, it also receives a third component of sunlight, which is reflected by
the surroundings, called albedo radiation; the reason that we can actually see the surroundings
is because these surroundings reflect part of the incoming visible light.
An illustration of the three components of sunlight is presented in Figure 3-1. These three
components of the radiation on the relevant surface sum up to the global radiation, as described
in Eq. (3-1).

20 Energy resources
Figure 3-1: The components of irradiance on a tilted surface.
Gglobal = Gdirect + Gdif f use + Galbedo (3-1)
The instantaneous amount of radiation that a surface receives is called irradiance, denoted
by G, and is expressed in [W m-2 ] or when multiplied with the surface area in [W]. The total
amount of radiation received by the surface over a certain time period - the integral of irra-
diance over time - is called irradiation, denoted by E, and is expressed in [Wh m-2 ] or [Wh]
or for long time periods in [kWh m-2 ] or [kWh].
3-2 Meteorological data
To make an accurate estimate of the available solar energy on a certain plane and location, the
global irradiance on that plane thus needs to be determined. Obviously the (global) irradi-
ance on that plane and location is varying over time. Day/night fluctuations exist due to the
rotation of the earth around its axis and seasonal fluctuations exist due to the earth’s axial tilt.
To make sure that these fluctuations are taken into account in the estimate of available solar
energy, historical measurement data on solar radiation can be used. The institute that has an
extensive database on solar radiation in the Netherlands (and a wide range of other ‘weather
parameters’) is the Royal Netherlands Meteorological Institute (KNMI). KNMI possesses a
weather station in the town of Cabauw, located in the province of South-Holland. The exact
location of this weather station is at 51.97o North (latitude), 4.93o East (longitude). Delft
has roughly the same latitude (52.00o N), but is located less then 40 km West from Cabauw,
at longitude 4.37o East. There could thus be a slight difference between the instantaneous
values of meteorological data in Delft and Cabauw; the distance between Delft and Cabauw

3-2 Meteorological data 21
Figure 3-2: Measuring solar irradiance [19].
however, corresponds to ±2 minutes in delay of the data, based on the rotational speed of the
earth. Furthermore, the local weather may differ between Delft and Cabauw as well. Both of
these effects are neglected in the processing of the meteorological data.
From the database of this weather station, the following data on solar radiation are used:
• Global Horizontal Irradiance (GHI): the global irradiance on a horizontal plane.

This is measured by a pyranometer: an instrument that translates the voltage that is
created by a thermopile detector into the associated irradiance level. The pyranometer
is mounted horizontally and has a field of view of 180o , the complete hemisphere, refer
#1 in Figure 3-2.
• Diffuse Horizontal Irradiance (DHI): the diffuse irradiance on a horizontal plane.

This is measured by a horizontally mounted pyranometer as well, but now a shade disk
blocks the direct irradiance from the sun, refer #2 in Figure 3-2. In this way, only the
diffuse component of the irradiance is measured.
• Direct Normal Irradiance (DN I): the direct irradiance on a plane normal to the irra-
diance vector.
This is measured using a pyrheliometer, an instrument with the same working principle
as a pyranometer, but with a limited field of view of only 5o . Mounted on a sun tracking
system, this instrument is always pointed towards the sun disk and in this way only
measures the direct component of the irradiance, refer #3 in Figure 3-2.
All of this data is recorded in intervals of 1 minute. For the calculation of the available solar
energy a total of 2 years (2012 & 2013) is taken into account.

22 Energy resources
(a) GHI in 2012 (b) GHI in 2013
Figure 3-3: Global Horizontal Irradiance in the years 2012 and 2013.
3-3 Irradiance on a horizontal plane
In this section, the irradiance data is discussed for the years 2012 and 2013. When the GHI
data is plotted over an entire year, presented in Figure 3-3, some obvious trends can be no-
ticed. During the middle of the year, corresponding to summer, the irradiance is a lot higher
than during the beginning and end of the year, corresponding to winter. The main reason for
this is the earth’s axial tilt; in summer the sun rays will fall more directly on the earth than
in winter. The larger the latitude of the location, the stronger this effect will be.
In Figure 3-4 the GHI and DN I are shown for two different days: one sunny day (May 28th
2013) and one cloudy day (May 29th 2013). As expected, the GHI peaks around noon when
the sun reaches the highest altitude of that day (also called solar noon) and goes to zero
during the night. This effect is caused by the rotation of the earth around its axis.
The second observation is that cloud cover has a great influence on the direct irradiance level,
shown in Figure 3-4a. On May 28th there is significant direct irradiance in contrary to May
29th , where there is no direct irradiance. This has a great influence on the global irradiance
as well, see Figure 3-4b. On the cloudy day only diffuse light is falling onto the pyranometer,
whereas on the sunny day both diffuse and direct light add up to the global irradiance.
Figure 3-3 shows that the pattern of irradiance is similar for the years 2012 and 2013. This
is in agreement with the findings of Narayan [20] on the irradiance in the years 2009-2012.
3-3-1 Irradiation on a horizontal plane
To make a quantitative comparison between the years 2012 and 2013, the yearly irradiation
on a horizontal surface can be calculated from the data, as described by Eq. (3-2), Eq. (3-3)
and Eq. (3-4). As it concerns a horizontal surface, the albedo irradiance is assumed zero;
there is no irradiance reflected by the surroundings onto the surface.

3-3 Irradiance on a horizontal plane 23
(a) Direct Normal Irradiance. (b) Global Horizontal Irradiance
Figure 3-4: Irradiance levels on a sunny and a cloudy day.
2012 2013
Eglobal,horizontal [kWh m-2 ] 1052.7 1057.4
Edif f use,horizontal [kWh m-2 ] 603.0 602.1
Edirect,horizontal [kWh m-2 ] 449.7 455.3
Table 3-1: Irradiation on a horizontal surface in 2012 and 2013.
Z
Eglobal,horizontal = GHI(t) · dt (3-2)
Z
Edif f use,horizontal = DHI(t) · dt (3-3)
Edirect,horizontal = Eglobal,horizontal − Edif f use,horizontal (3-4)

where:
GHI(t) = the instantaneous global horizontal irradiance [W m-2 ],
DHI(t) = the instantaneous diffuse horizontal irradiance [W m-2 ],
Eglobal,horizontal = the global horizontal irradiation [kWh m-2 ],
Edif f use,horizontal = the diffuse horizontal irradiation [kWh m-2 ],
Edirect,horizontal = the direct horizontal irradiation [kWh m-2 ].
The results for the horizontal irradiation in 2012 and 2013 are presented in Table 3-1.
The yearly global irradiation values for a horizontal surface of the years 2012 and 2013 are
thus very similar. In fact, taken also the data from [20], the deviation between the smallest
and largest values in those 5 years (1023 kWh m-2 and 1072 kWh m-2 ) is still less than 5%.
The assumption can now be made that the irradiance data of the years 2012 and 2013 are
representative for the irradiance in the near future.
Another observation can be made from Table 3-1, which is that the yearly diffuse irradiation
on a horizontal surface is larger than the direct irradiation. This is not always true for any

24 Energy resources
orientation of the surface; in fact, the tilt angle of the surface can have a big influence on
both irradiance components, as will be discussed in section 3-4.
As was concluded for the global irradiation, both of the direct and diffuse components don’t
vary much between the two years either: 0.1% difference in annual diffuse irradiation and
1.2% for the direct component.
3-4 Irradiance on a tilted plane
For the infotainment spot, where the PV module will be mounted on the back side, it is
necessary to make an accurate estimation of the irradiance on that part of the product. As
this back side is not horizontal, the irradiance on a tilted plane should be calculated, which
can be done with the measurement data from KNMI as well. In order to do so, all three
components of irradiance on a tilted plane - direct, diffuse and albedo - will be calculated
separately and consequently added together. The following sections describe how each of the
components is calculated.
3-4-1 Direct irradiance
The direct irradiance on a tilted plane can be calculated from the DN I; it is the fraction of
the DN I vector that is directed normal to the tilted plane of interest. In other words: it is
the DN I vector multiplied with the cosine of the angle of incidence (AOI), refer Eq. (3-5).
Gdirect,tilted = DN I · cos(AOI) (3-5)

where:
Gdirect,tilted = the direct irradiance on the tilted plane [W m-2 ],
DN I = the direct normal irradiance [W m-2 ],
AOI = the angle of incidence [rad].
Angle of incidence
To define the AOI, some parameters need to be defined first: the azimuth angle and the
altitude angle of an object (e.g. the sun) in space, see Figure 3-5a. The azimuth angle of the
sun (Asun ) is defined as the angle that the sun makes with respect to North, when seen from
above. The altitude angle of the sun (asun ) is defined as the elevation angle of the sun above
the horizon.
For clarity also the plane of interest is drawn in Figure 3-5b. The azimuth angle of the
plane (Aplane ) is defined in the same way as Asun , the angle with respect to North; the tilt
angle of the plane (θplane ) is defined as the angle that the plane makes with the horizontal
plane.
The AOI is then defined with trigonometry as:
AOI = cos−1 · [sin(asun ) · cos(θplane ) + cos(asun ) · sin(θplane ) · cos(Aplane − Asun )] (3-6)

3-4 Irradiance on a tilted plane 25
(a) Definition of Asun and asun . (b) Definition of Aplane and θplane .
Figure 3-5: Definition of azimuth, altitude and tilt angle.
where:
AOI = the angle of incidence [rad],
asun = the altitude angle of the sun [rad],
θplane = the tilt angle of the plane [rad],
Aplane = the azimuth angle of the plane [rad],
Asun = the azimuth angle of the sun [rad].
When Eq. (3-6) is substituted into Eq. (3-5), the equation to calculate the direct irradiance
on a tilted plane arises:
Gdirect,tilted = DN I · [sin(asun ) · cos(θplane ) + cos(asun ) · sin(θplane ) · cos(Aplane − Asun )] (3-7)
Sun Path Algorithm
When the plane of interest will have a fixed orientation, as is the case in the infotainment
spot, the variables θplane and Aplane in Eq. (3-7) will be constant. The azimuth angle of the
sun and altitude angle of the sun however are a function of time, as the sun moves across
the sky along the so-called sunpath. This sunpath depends on the location on earth, the time
of day and the time of year; the sunpath for Delft at different days throughout the year is
plotted in Figure 3-6. On December 21st the sun will have the lowest altitudes of the year and
also the shortest appearance above the horizon. On June 21st the sun will appear above the
horizon the longest and it will reach the highest altitude of the year. The sunpath of another
day, September 21st , is plotted as well for comparison.
To calculate the position of the sun in the sky, an Sun Path Algorithm (SPA) is used. Based
on the longitude and latitude of the location on earth an SPA calculates Asun and asun as
a function of time. The SPA developed by NREL [21] has been used to calculate Asun and
asun in minute resolution for the locations on the campus of TU Delft and consequently with

26 Energy resources
Figure 3-6: The sunpath for different days in the year.
Eq. (3-7) the direct irradiance vector is calculated for a range of orientations (θplane and
Aplane ). The results are presented in Figure 3-7.
As the location lays on the northern hemisphere of the earth, it is obvious that the optimal
azimuth angle is south. From the results the optimal azimuth angle is determined to be 181o ,
where the maximum direct yearly irradiance is found to be 590.6 kWh m-2 at a surface tilt
angle of 42o .
3-4-2 Diffuse irradiance
To calculate the diffuse irradiance on a tilted plane, several models have been developed, each
based on different principles. There is one thing that all these models have in common, which
is the sky view factor (SV F ). The SV F stands for the fraction of the sky that a surface
sees. In the ideal case, a horizontal surface without surrounding objects, this SV F equals 1,
i.e. the surface sees the complete hemisphere, as shown in Figure 3-8a. When a surface is
tilted however it sees a smaller fraction of the sky, depending on the tilt angle θ, as shown in
Figure 3-8b. When the tilt angle θ of the surface would be 90o , it would result in a SV F of
0.5. The relationship between the SV F and the tilt angle θ that is widely used in literature
is given in Eq. (3-8).
1 + cos(θ)
SV F = (3-8)
2
where:
SV F = the sky view factor,
θ = the tilt angle of the surface [rad].

Figure 3-7: Direct irradiation on a tilted plane [kWh m-2 yr-1 ].
(a) SV F of a horizontal surface. (b) SV F of a tilted surface.
Figure 3-8: The influence of tilt on the SV F , adapted from [22].

28 Energy resources
Sky Diffuse Models
The SV F is used in the calculation of Gdif f use,tilted in all models that are developed. These
models differ however from each other in complexity. In this section four models will be
discussed and a comparison between them will be made.
Isotropic Sky
The isotropic sky diffuse model is the most simplistic of the four models and forms the
foundation upon which more complex models are built. As the name reveals, it assumes that
the sky is isotropic over the entire hemisphere, meaning that the irradiance from each point
of the hemisphere falling onto the tilted surface is equal. The relationship between the diffuse
irradiance on the tilted surface and the SV F is presented in Eq. (3-9).
Gdif f use,tilted = DHI · SV F (3-9)
Sandia
David King developed an empirical model at Sandia National Laboratories [23]. It adds an
empirical correction term to the diffuse irradiance estimated by the isotropic sky model, to
account for the circumsolar and horizon brightening effects. The model is represented by
Eq. (3-10).
π

0.12 · 2 − asun − 0.04 · (1 − cos(θ))
Gdif f use,tilted = DHI · SV F + GHI · (3-10)
2
which can be rewritten as:
π

Gdif f use,tilted = DHI · SV F + GHI · 0.12 · − asun − 0.04 · (1 − SV F ) (3-11)
2
Hay & Davies

The Hay and Davies model divides the sky diffuse irradiance into an isotropic component and
a circumsolar one; horizon brightening effects are not included. An anisotropy index, Ai , is
defined in the following way:
DN I
Ai = (3-12)
G0
where:
DN I = the direct normal irradiance,
G0 = the extraterrestrial irradiance.
The diffuse irradiance on the tilted plane is then modeled according to Eq. (3-13):
Gdif f use,tilted = DHI · [Ai · cos(AOI) + (1 − Ai ) · SV F ] (3-13)

Figure 3-9: Diffuse irradiation on a tilted plane [kWh m-2 yr-1 ].
Reindl
The Reindl sky diffuse irradiance model [24] [25] [26] represents three components of diffuse
radiation on the POA, including isotropic, circumsolar brightening, and horizon brightening.
This model extends the Hay and Davies model by adding an additional factor to the ‘bright-
ening’ term to account for horizon brightening. The formulation of the sky diffuse irradiance
is then equated as follows:
  s 
DN I · cos(θ) θ 

Gdif f use,tilted = DHI · Ai · cos(AOI) + (1 − Ai ) · SV F · 1 + · sin3
GHI 2
(3-14)
Comparison
The yearly diffuse irradiation on a tilted plane, estimated by the above-mentioned models
is plotted in Figure 3-9 as function of the tilt angle of the surface. The isotropic sky model
decreases the most with increasing tilt angle, as this model does not take horizon brightening
or circumsolar brightening into account. The Sandia model does take both into account and
therefore the irradiation plot decreases less with increasing tilt angle.
It is remarkable that both the Hay & Davies and the Reindl model forecast a lower irradiation
at lower tilt angles, although from the DHI data is calculated that in the year 2013 - which
data is also used for this calculation - 602.1 kWh m-2 is irradiated on a horizontal surface,
refer Table 3-1. Both of them are less dependent on the tilt angle than the isotropic or Sandia
model, even though the Hay & Davies model does not account for horizon brightening effects.
Advanced sky view factor
In the case of the infotainment spot a more detailed estimation of the diffuse irradiation has
to be performed. The reason for this is that the diffuse sky models in 3-4-2 assume that the

30 Energy resources
Figure 3-10: Sky View Factor as function of tilt angle.
SV F is only affected by the tilt angle of the surface itself, which is not the case. On the two
proposed locations a fraction of the sky will also by blocked by the surroundings, especially
by the CEG and EEMCS buildings.
To determine a more realistic SV F that takes the surroundings into account as well, a 3D
modeling approach is chosen. In Google SketchUp the location of the infotainment spot is
modeled including its surroundings. The longitude and latitude of the locations are used
as inputs and the models of the EEMCS and CEG buildings were available in the so-called
Extension Warehouse. The trees that are in the vicinity of the proposed locations have
been added in the 3D model as well. A picture of the created model is shown in section
3-4-4, in Figure 3-14a. On the proposed locations a tilted surface is created and for that
surface the SV F is calculated by a plugin: LSS Chronolux. The azimuth angle of the surface
is chosen to be 180o , as this is the proposed azimuth angle for the infotainment spot. The
tilt angle of the surface is varied from 0o to 90o and for each tilt angle the SV F is determined.
The SV F determined from Google SketchUp is plotted in Figure 3-10 together with the SV F
that is used in the sky diffuse models. Obviously the SV F that is calculated with both the
tilt angle and the surroundings taken into account is smaller than the SV F based only on the
tilt angle. The blue area corresponds to the fraction of the hemispherical sky that is blocked
due to the tilt angle of the surface; the red area represents the fraction of the sky blocked
by the surroundings and the green area is the fraction of the sky that is still ‘visible’ for the
tilted surface.
When the SV F from the models in section 3-4-2 is replaced by the SV F obtained by 3D-
modeling, the diffuse irradiation can be calculated more accurately and the results are pre-
sented in Figure 3-11. Again, the isotropic sky model has the strongest dependency on the
tilt angle for the reasons mentioned in 3-4-2 and is the most pessimistic for all tilt angles.
The Sandia model claims the same irradiation at a horizontal surface, but is more optimistic
for higher tilt angles, as explained earlier. For medium and high tilt angles, the Sandia model

Figure 3-11: Diffuse irradiation on a tilted plane [kWh m-2 yr-1 ], location-specific SV F .
predicts a diffuse irradiation very close to the value that both the Hay & Davies model and
the Reindl model are predicting. It is remarkable that the latter two models predict a higher
horizontal irradiation, in contrast to the results in Figure 3-9.
Based on these results, it is chosen to perform further calculations with the Sandia model. In
combination with the location-specific SV F it shows a close match with the isotropic model
for small tilt angles; for larger tilt angles it approaches the results of the other models by an
empirical correction for horizon brightening effects and circumsolar brightening.
3-4-3 Albedo irradiance
The third and last irradiance component on a tilted surface is originated from the reflected
irradiance off the ground. This albedo irradiance is thus a function of the reflectivity of the
ground surface, known as albedo, and the tilt angle of the surface.
1 − cos(θ)
Galbedo,tilted = GHI · albedo · [27] (3-15)
2
which can be rewritten as:
Galbedo,tilted = GHI · albedo · (1 − SV F ) (3-16)
The reflectivity of the ground surface depends on the ground type; typical values for various
surface types are presented in Table 3-2.
The tabulated albedo values give an indication of some widely used surface types. For a more
accurate albedo estimate on a specific location, one could perform measurements on-site. The

32 Energy resources
Surface type Albedo

Grass (summer in UK) 0.25
Lawns 0.18-0.23
Dry grass 0.28-0.32
Bare soil 0.17
Asphalt 0.15
Concrete 0.20
Table 3-2: Typical albedo values for various ground types [28].
(a) Measuring albedo of the grass. (b) Measuring albedo of the slabs.
Figure 3-12: Setup for albedo measurements on campus grounds [29].
actual albedo of the campus grounds is measured by a TU Delft student in the year 2014 [29].
The measurement setup consisted of an upward facing (horizontal) radiometer and a down-
ward facing one, see Figure 3-12. The first radiometer was used to measure the irradiance
level falling from above onto a horizontal surface at ±1.5 m height, i.e. the GHI. The latter
radiometer measured the irradiance on the same surface, but coming from the bottom, so
reflected by the ground. The albedo of the ground is consequently determined as the fraction
of the GHI that is reflected by the ground. This measurement was performed on the cam-
pus grass as well as the slabs that form the footpaths on campus. Results are presented in
Table 3-3.
Most of the ground type that the photovoltaic (PV) module on the back side of the infotain-
ment spot faces will be grass. To make a correction for the parts that are not grass an average
albedo of 0.20 is assumed. This value for albedo is widely used for a general estimate of the
albedo irradiance [28].
grass slabs
GHI [W m-2 ] 185.0 200.6
Gref lected [W m-2 ] 53.8 26.3
albedo 0.29 0.13
Table 3-3: Albedo on TU Delft campus [29].

Figure 3-13: Albedo irradiation on a tilted plane [kWh m-2 yr-1 ].
The prediction of the yearly albedo irradiation is presented in Figure 3-13. Obviously this
irradiation component increases with increasing tilt angle, as the inclined surface then sees a
larger part of the ground surface, where the albedo irradiation is reflected from. The irradia-
tion level of the location-specific SV F is higher than for the original SV F , because the SV F
itself is lower. In other words: the tilted plane sees not only the ground surface, but also the
surrounding objects, such as trees and buildings, where albedo irradiation can originate from.
Intermezzo
The location-specific SV F is determined with the goal to make a more accurate estimate
on the solar energy availability. The SV F is lower than the one from literature as it takes
the effect of the surroundings into account on top of the tilt angle dependency. This lower
value for SV F has a negative effect on the diffuse irradiation, but a positive effect on the
albedo irradiation. The question arises if both effects cancel each other out or not. If so, the
location-specific SV F has no additional value in determination of the global yearly irradiation.
However it would still be more accurate in the determination of the different components of
the yearly irradiation, i.e. diffuse and albedo irradiation. After the following section, where
shading is discussed, a conclusion is drawn on the use of the location-specific SV F .
3-4-4 Shading
As briefly touched upon in section 2-3-2, shading will influence the irradiance on a surface.
Shading of the sun can be caused by clouds in the atmosphere or by the surroundings of the
(tilted) surface. Shading by means of clouds has been taken into account in the calculation,
as this is already implemented in the DN I vector. At times the sun is blocked by the sur-
roundings however, the direct irradiance will be zero as well. To account for this loss in direct
irradiance, two methods have been used: a 3D modeling approach and one with measure-
ments on-site. Both methods are explained in the next sections, followed by a concluding
comparison between them.

34 Energy resources
(a) Overview of the surroundings. (b) The infotainment spot itself.
Figure 3-14: The infotainment spot modeled in Google SketchUp.
Google SketchUp
The 3D model that has been made in Google SketchUp to determine the location-specific
SV F has also been used to determine a shading coefficient. The model is shown in Figure 3-
14.
Within SketchUp a SPA is incorporated and thus it can be used to simulate the shadow that
objects create, as the longitude and latitude of the location are entered as well. In this way,
for each day in the year the shading on the infotainment spot induced by its surroundings is
simulated. In general it was found that both proposed locations suffer from shading by the
Civil Engineering and Geosciences (CEG) building early in the morning and by the Electrical
Engineering, Mathematics and Computer Science (EEMCS) building in the afternoon. Lo-
cation 1 is shaded sooner in the afternoon than location 2, but for location 1 the shading by
the EEMCS building disappears at some point in the late afternoon, whereas for location 2
this does not happen.
An overview of the shading analysis on August 12th 2014 on location 2 is shown in Figure 3-
15. In the same figure the validation of the model is presented. It shows that the SketchUp
prediction at this date was very accurate with respect to the reality. In Figure 3-15e the
shading coefficient that is created for this day is presented. Shading coefficient equal to 1
means that the infotainment spot is shaded; 0 means that there is no shading. The time that
the shading factor changes is different each day and is determined to the minute.
The shading coefficient that is created in this way affects only the direct component of the
irradiance at the location; the instantaneous direct component is calculated using Eq. (3-7)
and consequently multiplied with (1 - the instantaneous shading coefficient). The diffuse and
albedo component are calculated using the location-specific SV F to determine the global
irradiance and irradiation on a tilted plane. The results are presented after the next section,
where another way of determining the shading coefficient is described.

(a) 10:25 in Reality. (b) 10:25 in SketchUp.
(c) 16:58 in Reality. (d) 16:58 in SketchUp.
(e) Shading coefficient on August 12th 2014.
Figure 3-15: Creating the shading coefficient on August 12th 2014 using SketchUp.

36 Energy resources
Meteonorm Horicatcher
Another way of determining the shading coefficient by the surroundings is done by means of
the Horicatcher. The Horicatcher is a tool created by Meteonorm with the purpose to map
the horizon as seen by a PV module. It consists of a reflecting hemisphere which can be
placed at the location where a PV module is to be installed. With a digital photo camera
a picture is taken from straight above the hemisphere. The software that is provided with
the tool is able to convert this photo into a picture of the horizon that the PV module faces.
Consequently the objects that would block the sun can be selected manually and their opacity
can be set.
The software Meteonorm then makes use of satellite data to calculate the irradiance on a
(tilted) surface at that location for each hour of the day throughout the year. In the case of
the infotainment spot this feature is used to determine the (hourly) shading coefficient. The
global irradiance on the tilted surface is calculated with the true horizon, with the shading by
the buildings and trees on campus. Shading that is caused by trees is incorporated into the
horizon for the months Mar-Oct, whereas for Nov-Feb the trees are assumed leafless. Next,
the global irradiance on the tilted surface is calculated for the same location, but now without
any horizon, meaning that no buildings or trees would shade the (tilted) surface. The shading
coefficient that is solely caused by the surrounding objects (so not by the tilt angle of the
surface itself) is then calculated, refer Eq. (3-17).
irradiance with horizon

shading coef f icient = (3-17)
irradiance without horizon
The shading coefficient that is calculated in this way affects thus the global irradiance on a
tilted plane. The direct component of irradiance is therefore calculated with Eq. (3-7) and
for the diffuse en albedo irradiance the SV F of the models - only a function of tilt angle - is
used. This is done because the losses in diffuse and the gain in albedo irradiance is already
incorporated in the shading coefficient. Consequently the three components are summed up
to the global irradiance, which is then multiplied with the shading coefficient.
Comparison
As explained, the shading coefficients are determined in two different ways and affect different
parameters in the calculation of the irradiance (and irradiation) on a tilted plane. It would
therefore not make sense to compare the shading coefficients themselves; instead the irradi-
ance during a day is compared in both situations, refer Figure 3-16. As can been seen, the
daily profile of global tilted irradiance differs for both models. The rise in the irradiance level
around 10 AM modeled by SketchUp is at the time that the infotainment spot comes out
of the shade caused by the CEG building. A similar rise in global tilted irradiance modeled
by Meteonorm is seen roughly 1 hour later. Also the irradiance level predicted by SketchUp
is changing more abrupt, as the direct component gets completely blocked or unblocked,
which makes it a more realistic daily profile than the one predicted by Meteonorm, where the
changes in irradiance level is more gradual.

Figure 3-16: Global irradiance on a tilted plane for two different shading models.
Irradiation [kWh m-2 ] Meteonorm SketchUp Deviation

2012 - Location 1 805.8 800.1 1.1%
2012 - Location 2 835.3 819.4 2.6%
2013 - Location 1 815.5 806.7 0.7%
2013 - Location 2 840.6 819.5 1.9%
Table 3-4: Yearly irradiation predicted with two shading models.
To see what the difference between the shading coefficients is in terms of irradiation, the
cumulative irradiance for location 1 in the year 2013 is plotted in Figure 3-17 and the yearly
irradiation is presented in Table 3-4 as well.
From Figure 3-17 it is obvious that both shading coefficients have almost the same effect on
irradiation, even though the daily profiles may differ. From Table 3-4 it shows that this holds
true for both locations in both investigated years; the biggest deviation in yearly irradiation
between the two shading analyses is found in 2012 for location 2: 2.6%. The conclusion can
be drawn that both shading analyses deliver similar results in terms of irradiation, but the
SketchUp analysis results in a more realistic daily irradiance profile.
3-4-5 Summation of the components
The three components of irradiance on a tilted plane have been determined in the previous
sections. The global irradiance is then calculated by simply summing up the three compo-
nents. In Table 3-5 the results are presented. For the diffuse component in this calculation
the Sandia diffuse sky model is used and the shading is simulated with the SketchUp shad-
ing coefficients affecting the direct component. Furthermore, the location specific SV F and
shading coefficient of location 1 are used.

38 Energy resources
Figure 3-17: Cumulative global irradiation on a tilted plane for two different shading models.
Global Direct Diffuse Albedo Tilt Azimuth

horizontal no surroundings 1052.2 450.1 602.1 0 - -
horizontal location 1 807.1 332.3 405.8 68.9 - -
tilted no surroundings 1156.0 581.7 559.1 15.1 31o 182o
tilted location 1 869.1 433.4 345.7 90.0 33o 183o
Table 3-5: Irradiation in 2013 [kWh m-2 yr-1 ].

A number of observations and conclusions can be drawn from this table of results:
• The surroundings where the surface is placed has a huge impact on the irradiance that
the surface receives. The maximum global irradiation on a horizontal surface for this
location decreases 23.3% due to the surroundings; for an optimally tilted this decrease
is 24.8%.
• The relative decrease in diffuse irradiance due to the surroundings is more than that
of the direct component. Where the direct component decreases 26.2% and 25.5% for
a horizontal and optimally tilted surface respectively, the diffuse component decreases
32.6% and 38.2%.
• Note that the direct component on a horizontal surface - calculated with Eq. (3-7) -
is very close to the value that was obtained earlier in Eq. (3-4): 450.1 and 455.3 kWh
m-2 yr-1 respectively, meaning a 1.2% deviation. This means that the approach of
multiplying the DN I vector with the cosine of the AOI gives an accurate estimate of
the direct component on a (tilted) surface.
The global irradiation for the (tilted) surface in both cases (without surroundings and with
surroundings) are plotted as function of azimuth and tilt angle in Figure 3-18 and Figure 3-19
respectively.
From Figure 3-18 it can be seen that the global irradiation profile looks very similar to that of
the direct irradiation profile from Figure 3-7, but shifted down in terms of tilt angle. This can
be explained by the fact that diffuse irradiation is a big contributor to the global irradiation
and the diffuse irradiation is larger for smaller tilt angles. The fact that albedo irradiation is
larger for larger tilt angles does not weigh up to this effect, as its contribution to the global
irradiation is far smaller.
Figure 3-18: Global irradiation on a tilted plane without surroundings [kWh m-2 yr-1 ].

40 Energy resources
Figure 3-19: Global irradiation on a tilted plane on location 1 [kWh m-2 yr-1 ].
From Figure 3-19 another observation can be made; the contourplot is not symmetrical in
terms of azimuth angle, whereas the contourplot of Figure 3-18 is. For higher tilt angles it
would benefit to have an azimuth angle smaller than 180o . At an azimuth angle of 180o the
surface is directed slightly towards the EEMCS building; if the surface would have a smaller
azimuth angle, the AOI during the time that the surface is not yet shaded by the EEMCS
building is smaller, resulting in a higher direct component of irradiance.

3-5 Conclusion 41
Intermezzo
The question if the location-specific SV F has a big impact can now be treated. The location-
specific SV F is larger than the SV F used in the models, resulting in a smaller diffuse irra-
diance component, but a larger albedo component. To see whether or not both effects cancel
each other out, the diffuse irradiation plus the albedo irradiation is plotted as function of the
tilt angle, for both sky view factors in Figure 3-20.
Figure 3-20: Diffuse irradiation + Albedo irradiation for different SV F s.
From Figure 3-20 it can clearly be seen that the location-specific SV F has an impact on the
yearly irradiation, irrespective of the tilt angle of the surface. Therefore it can be concluded
that an irradiation calculation should be done with a SV F that is specific for the particular
location in order for it to be more accurate.
3-5 Conclusion
The irradiance on a (tilted) plane consists of three components: direct, diffuse and albedo
irradiance. Different diffuse sky models have been looked into, the influence of a location-
specific SV F has been shown and the influence of shading is performed in two different ways.
Altogether it shows from each calculation that the azimuth angle that results in the highest
irradiance on a plane is South. Therefore, from this point on, the azimuth angle that is
suggested for the infotainment spot is 180o and this is used in the further calculations. The
individual components of the infotainment spot will be calculated in the following way:
• The direct component is calculated by multiplying the DN I with the AOI and accord-
ingly with (1-shadingfactor) determined by Google SketchUp.
• The diffuse component is calculated by the Sandia diffuse sky model in combination
with the location-specific SV F , determined by Google SketchUp.
• The albedo component is calculated by multiplying the GHI with (1-SV F ), again with
the location-specific SV F .

42 Energy resources

Chapter 4
PV Potential
In the previous chapter, the irradiance and irradiation on a surface with a certain tilt and
azimuth on the proposed locations of the infotainment spot are determined. In this chapter,
this methodology is applied to the surface of the infotainment spot on which the photovoltaic
(PV) module will be mounted. This surface, however, is not only tilted, but is also curved.
This makes it more complex to determine the irradiance incident on it. In section 4-1 this
issue will be discussed together with the method that is used to overcome the issue. In section
4-4 a variety of flexible PV modules - available on the market today - will be introduced.
Furthermore, the performance of these modules under the irradiance conditions as determined
in 4-1 will be discussed.
4-1 Curvature
In the design considerations it is decided that the flexibility of the PV module should be
used in the design. This has resulted in the fact that the PV module is to be mounted on a
curved surface on the back of the infotainment spot. This implies an additional complexity
in determining the irradiance on the PV module; the azimuth angle will be constant for the
entire module, but the tilt angle is varying. In the following section 4-1-1 a methodology will
be discussed to tackle this problem.
4-1-1 Methodology
In order to accurately estimate the irradiance on a tilted surface, the tilt angle should be
known. Therefore, a method is chosen to divide the curved plane into multiple flat segments,
each with a fixed tilt angle. Obviously, the larger the amount of segments used, the better
it would mimic the reality. However, the computational time will be longer. Furthermore,
as each segment would have a different location-specific sky view factor (SV F ), it would be
preferable that the amount of segments is small, so that the determination of the location-
specific SV F would be less time consuming. The following question thus arises:

44 PV Potential
What is the minimum amount of segments that would give an accurate representation of the
curved surface?
To answer this question, the yearly irradiation for a variety of segments is calculated and is
judged on the following criteria:
• The percentage of increase in yearly irradiation by adding an additional segment, re-

ferred to as ‘marginal irradiation’.
• The percentage deviation of the yearly irradiance from that of the case with 104 seg-
ments, which is assumed to be a highly accurate representation of the reality.
In Figure 4-1 a side-view of part of the infotainment spot is shown, namely the surface where
the PV module is to be mounted. The length of the considered flexible PV modules (refer
4-4) differs from module to module, which means that some flexibility in the dimensions of
the curvature is favorable.
The height of the highest point of the curvature (H) is determined from the desired screen
height, the width of the screen and the width of the passe partout (the area around the screen)
in Eq. (4-1). The width of the passe partout is at this stage set to half the screen width.
1
H = Hscreen + · wscreen · sin(θscreen ) + wpp · sin(θscreen ) (4-1)
2
H = 1.25 + 0.16 · sin(65o ) = 1.40 (4-2)
where:
H = the height of the highest point of the curvature [m],
Hscreen = the height of the middle of the screen [m], as determined in Chapter 2,
wscreen = the width of the screen (not of the entire tablet) [m],
wpp = the width of the passe partout [m].
The flexibility of in the dimensions of the curvature is implemented by a variable footprint (F ).

Depending on the length of the PV module, the F can be chosen such that the length of the
curvature (L) is large enough for the PV module to be placed on the curvature. The curvature
is part of a circle with radius R and with its center C at height H. The height of the center
C is fixed, however the radius R depends on the footprint that is chosen. The last parameter
to introduce is α, which is the angle that the curvature makes from top to bottom.
This model has two limitations:
1. R ≥ 1.40 m, as the curvature would otherwise not reach the ground;
2. 0 ≤ F ≤ 1.40 m, so that the curvature describes any situation between a flat vertical
surface up to a quarter of a circle.

4-1 Curvature 45
Figure 4-1: Definition of variables.
Mathematical model
From Figure 4-1 the following formula for radius R can be derived:
F = R (1 − cos(α)) (4-3)
F
R= (4-4)
1 − cos(α)
Radius R can also be described by other parameters from Figure 4-1:
H
R= (4-5)
sin(α)
When Eq. (4-4) is substituted into Eq. (4-5), a formula for the angle α can be found:
F H
= (4-6)
1 − cos(α) sin(α)
F 1 − cos(α) α

= = tan (4-7)
H sin(α) 2
F

−1
α = 2 · tan (4-8)
H
The angle α is used to divide the the curvature into segments of equal size and it forms the
basis of the calculation of the tilt angles of those segments. Figure 4-2 shows the division of
the curvature into 1, 2, 3 and 4 segments in 4-2a, 4-2b, 4-2c and 4-2d respectively.
The angle that each segment makes, αs , is a function of the amount of segments n:
α
αs = (4-9)
n

46 PV Potential
(a) 1 segment (b) 2 segments
(c) 3 segments (d) 4 segments
Figure 4-2: Division of the curved plane into flat segments.

4-1 Curvature 47
Figure 4-3: The top and bottom angle of segment i.
Next, the angle that the top and bottom of segment i (refer Figure 4-3) make can be deter-
mined:
αi,top = (i − 1) · αs for 1 ≤ i ≤ n. (4-10)

αi,bottom = i · αs for 1 ≤ i ≤ n. (4-11)
The tilt angle of segment i can then be described as function of αi,top and αi,bottom as follows:
!
−1 R · sin(αi,bottom ) − R · sin(αi,top )
θi = tan (4-12)
R · cos(αi,top ) − R · cos(αi,bottom )
Now the tilt angle for each segment is known, the irradiance and irradiation on this segment
can be calculated.
Gi,global (t) = Gi,direct (t, θi , Ai , asun , Asun ) + Gi,dif f use (t, SV F ) + Gi,albedo (t, SV F, albedo)
Z (4-13)
Ei,global = Gi,global (t)dt (4-14)
The average irradiance and average irradiation falling onto the surface is then computed as
follows:
1 i=n
X
Gaverage,global (t) = Gi,global (t) (4-15)
n i=1
1 i=n
X Z
Eaverage,global = Ei,global = Gaverage,global (t)dt (4-16)
n i=1
For this calculation the largest footprint is chosen: F = 1.40 m. If a certain amount of
segments for this footprint would describe the curvature accurate enough, a smaller footprint

48 PV Potential
would not pose any problem: the curvature will have a bigger radius, so will be even more
accurately described by the same amount of segments. Furthermore, an azimuth angle Ai of
180o is chosen, for this is (approximately) the desired azimuth angle of the infotainment spot.
Eq. (3-7) is used to calculate Ei,direct , which is consequently multiplied with the shading
coefficient obtained by SketchUp. Ei,dif f use is calculated with the Sandia diffuse sky model,
Eq. (3-10), with the SV F based only on the tilt angle of the segment θi . This is chosen,
because otherwise for each configuration for each segment a SV F had to be determined,
while in this calculation the magnitude of Ei,dif f use by itself is not of utmost importance,
as it will be compared with the same parameter under other configurations. Ei,albedo is
determined with Eq. (3-16), again with the SV F based only on the tilt angle of the segment
θi and albedo considered 0.20.
4-1-2 Results
Eaverage,global is calculated in this way for different amounts of segments. To make a valid
comparison between the different configurations, the irradiation is multiplied with the total
length of the surface, resulting in irradiation per width of the surface [kWh m-1 yr -1 ], referred
0
to as Eaverage,global :
0
Eaverage,global = Eaverage,global · n · lengthsegment (4-17)
with q
lengthsegment = (R · (1 − cos(αs )))2 + (R · sin(αs ))2 (4-18)
Results of a series of calculations are plotted in Figure 4-4 and Figure 4-5. The former figure
0
shows the percentage of accuracy improvement in Eaverage,global by adding 1 segment, with
0 0
respect to the previous value of Eaverage,global . This additional accuracy in Eaverage,global is
0
referred to as ‘marginal irradiation’. The latter figure shows the deviation in Eaverage,global
from when 104 segments are used.
From this series of calculations it can be concluded that only 4 straight segments would
already mimic the curvature accurately: the marginal yearly irradiation is with 0.23% far
below 1.0% and the deviation from the case with 104 segments is minimal with 0.29%. When
a larger amount of segments is chosen, the accuracy will increase even further, however it
would also increase the computational time for the following calculations and for determining
the location-specific SV F for all segments. For this reason, for further calculations 4 straight
segments have been used to describe the curvature.
Intermezzo
Earlier the claim is made that for smaller footprints the accuracy would be higher for a
0
fixed amount of segments. Figure 4-6 supports this claim: Eaverage,global is calculated with 4
segments for a variety of footprints. The deviation with respect to the case with 104 segments
is plotted versus the footprint.

4-1 Curvature 49
Figure 4-4: The marginal yearly irradiation on the surface.
Figure 4-5: The deviation in Eaverage,global

0
compared with 104 segments as function of n for
F =1.40 m.

50 PV Potential
Figure 4-6: The deviation in Eaverage,global

0
compared with 104 segments as function of F , for
n=4.
4-1-3 Available energy
With the amount of segments set to 4, the irradiance on the PV module on the back of
the infotainment spot can now be computed accurately. The model from Figure 4-1 is first
adapted to a more realistic one, where the PV module does not occupy the entire length of
the curvature, but some space on top and bottom is left open. For similar reasoning as with
the smaller footprints, the accuracy of simulating the curved surface is slightly increasing by
this adaptation of the model. The angle that the offsets on top and bottom of the module
make are referred to as β1 and β2 respectively, as shown in Figure 4-7.
The angle that each segment makes, αs , is now not only a function of the amount of segments
n, but the angles β1 and β2 as well. Also the top and bottom angle of each segment are now
depending on β1 , see Eq. (4-19) - Eq. (4-21).
α − β1 − β2
αs = (4-19)
n
αi,top = (i − 1) · αs + β1 for 1 ≤ i ≤ n. (4-20)
αi,bottom = i · αs + β1 for 1 ≤ i ≤ n. (4-21)
The irradiance on the module is consequently calculated with Eq. (4-15) and the irradiation
with Eq. (4-16) for a variety of footprints F and azimuth angles Amodule . The diffuse and
albedo component are just for this calculation computed with the SV F based on the tilt
angle of the segment. The determination of the location-specific SV F would otherwise be
unnecessarily time-consuming, as it is a function of both footprint F and azimuth A. The
location-specific SV F will be used to improve the accuracy of the computed irradiance vector
Gaverage,global (t) once a footprint has been chosen. The choice for a certain footprint will

4-1 Curvature 51
Figure 4-7: The offset margins on top of and below the module.
Figure 4-8: Eaverage,global as function of F and Amodule [kWh m-2 yr-1 ].
depend on the type of PV module that will be used as well (refer section 4-4) as the energy
demand by the electric system (refer chapter 5).
For this computation both the offsettop and offsetbottom are set to 0.10 m. The results are
plotted in Figure 4-8.
From Figure 4-8 it is obvious that the total annual irradiation is highest for Amodule close to
South and for the largest footprint F of 1.40 m. This agrees with the conclusions in chapter 3.
Another observation is that the footprint has a quite heavy influence on the annual irradiation;
a configuration with 1.40 m footprint with 90o or 270o azimuth angle (facing East or West
respectively) has higher annual irradiation than a configuration with 0.10 m footprint on an
180o azimuth angle.

52 PV Potential
Individual segments
Important to realize is that the irradiance vector on the module and the annual irradiation is
an average taken over the 4 segments. In this section, the difference between the irradiance on
the individual segments is discussed. In Figure 4-9, the irradiance on the individual segments
is plotted for several days throughout the year.
From Figure 4-9a it can be seen that when the direct component reaches the curvature, there
is quite some difference between the irradiance levels on the different segments. Segment 1
- the top segment with the largest tilt angle - has the highest irradiance, because the angle
of incidence (AOI) on this segment is smallest due to the low altitude of the sun in winter
(refer Figure 3-6). For the same reason, segment 4, the bottom segment with the smallest tilt
angle, receives the lowest irradiance during that day.
In summer, the situation is opposite as illustrated by Figure 4-9b. Due to the relatively high
position of the sun in the sky, segment 1 now has the largest AOI and therefore the lowest
irradiance. The bottom half of the curvature, segments 3 and 4, now experiences the highest
irradiance levels.
Figure 4-9c shows a day somewhere between the two extremes of summer and winter. The
sunpath is such that the AOI for segment 1 and 4 is similar, resulting in a similar irradiance
level. During this day, segments 2 and 3 experience the highest irradiance.
The last figure, Figure 4-9d shows a day where the direct irradiance is 0 throughout the entire
day. It is clearly visible that - even though the sun has a very low sunpath on January 3rd
- difference in irradiance between the segments is small. The same can be seen from the
previous 3 graphs at times when the direct component does not reach the curvature. It can
thus be concluded that the difference in irradiance levels between the segments is large only
in the winter and summer months at times when direct sunlight hits the curvature.
There is a reason why this difference in irradiance levels is discussed. Some PV modules
are configured in a way that this difference in irradiance has a negative influence on the
power output. A schematic of such configuration is presented in Figure 4-10a. When in
such configuration the individual solar cells are connected in series, the current produced by
each cell should match. If some cells experience lower irradiance than others however, they
will produce less current and force the other cells to produce the same (lower) current. In
this way, not all cells produce as much power as they could have, based on the irradiance level.
Some PV modules have a different configuration, as schematically shown in Figure 4-10b.

Such PV modules would not face this problem, as each individual cell would face the same
total irradiance level. At certain parts of the cells the current production would be higher
than at other parts of the cell, but this would not pose a limitation on the current production
by series interconnection; the total current production of each cell would be the same.
For the infotainment spot, PV modules with both configurations are looked at and further
discussed in section 4-4.

4-1 Curvature 53
(a) Gi,global on December 10th 2013. (b) Gi,global on June 7th 2013.
(c) Gi,global on September 5th 2013. (d) Gi,global on January 3rd 2013.
Figure 4-9: Irradiance level on the different segments in different times of year.
(a) Configuration A, 1 cell marked in red.
(b) Configuration B, 1 cell marked in red.
Figure 4-10: Schematic PV module configurations.

54 PV Potential
4-2 PV output
Now that the irradiance on the PV module is known, the electrical power production of the
PV module is to be determined. This section elaborates on a PV model that estimates the
output of a PV module under varying surrounding conditions of temperature and irradiance.
Both the solar cell temperature and the irradiance on the solar cell have an impact on the
solar cell performance. To characterize solar cells therefore a standard has been created, the
Standard Test Conditions (STC): the solar cell temperature is 25o C while it is irradiated by
the AM1.5 solar spectrum, corresponding to an irradiance level of 1000 W m-2 . A number of
flexible PV modules that could be suitable for the infotainment spot is analyzed in terms of
electrical output over a period of time, as a function of temperature and irradiance.
4-2-1 The photovoltaic effect
This section describes the working principle of a solar cell, though very briefly and simplified.
When light is irradiant on a solar cell, part of it will be absorbed by the cell, which results in
the generation of electron-hole pairs. The electrons and holes are separated by the junction
of the different solar cell layers. Consequently, the separated charge carriers are collected
at the terminals of the junction. When a load is connected between the terminals of an
illuminated solar cell, a current flows through the load and due to the voltage drop at the
load, voltage (V ), it performs work. The net current per unit area, or current density (J), of
an illuminated ideal solar cell is described by the Shockley solar cell equation [28]:
qV

J(V ) = Jph − J0 exp −1 (4-22)
kT
in which q is the electronic charge (1.602·10-19 C), k is the Boltzmann’s constant (1.380·10-23 J
K-1 ), T is the temperature in Kelvin and J0 is the dark current density, given by equation:
DN DP

2
J0 = qni + (4-23)
LN NA LP ND
in which ni is the intrinsic charge carrier concentration, DN and DP the diffusion coeffi-
cients of electrons and holes respectively, LN and LP the diffusion length of electrons and
holes respectively and NA and ND the concentration of acceptor and donor atoms respectively.
Eq. (4-22) can be plotted as JV-curve and, when multiplied by the area of the solar cell, it
can be converted to IV-curve1 . When the terminals of the solar cell are connected without an
electric load in between, the output current is referred to as short circuit current (Isc ). This
point is the intersection of the IV-curve and the vertical axis (at V = 0).
When the external circuit connected to the solar cell is not closed, it means no current
is flowing; the voltage at the terminals of the solar cell in that case is called open circuit
voltage (Voc ). This point is the intersection of the IV-curve and the horizontal axis (at
1
The IV-curve of a real solar cell differs from the ideal one, as series resistance and shunt resistance slightly
alter the curve.

4-2 PV output 55
(a) Influence of irradiance on the IV-curve. (b) Influence of temperature on the IV-curve.
Figure 4-11: The effect of irradiance and temperature on the IV-curve of a solar cell.
I = 0). The relation for Voc can be easily determined from equation 4-22 by substituting 0
for J(V );
kT Jph

VOC = ln +1 (4-24)
q J0
Based on Eq. (4-22) and Eq. (4-23) and the IV-curve (refer Figure 4-11), the effect of irradiance
and temperature on the output of a PV cell is discussed in the next sections.
4-2-2 Irradiance dependence
The fluctuations in solar irradiance are thoroughly discussed in chapter 3. The irradiance on
a PV cell will influence its electrical output. Higher irradiance means that more photons are
incident on the solar cell and thus more electron-hole pairs will be generated. Therefore, Jph
(Eq. (4-22)) will be higher and the net current produced by the solar cell will be higher as
well; this is illustrated in Figure 4-11a.
4-2-3 Temperature dependence
The electrical output of a solar cell depends on the cell temperature as well. At first glance,
one could think that equation 4-24 would imply a rise in VOC with an increase in temperature,
but from Figure 4-11b it shows clearly that this is not the case; this is caused by the term J0
in the equation. From equation 4-23 it is known that J0 is proportional to ni 2 . The term ni
will increase with increasing temperature, as shown in equation 4-25, in which B is a constant
which is essentially independent of temperature and EG0 is the band gap linearly extrapolated
to absolute zero [30]. This increase in ni causes a big increase in J0 , which in turn causes a
decrease in VOC with increasing temperature.
EG0

ni 2 = BT 3 exp − (4-25)
kT

56 PV Potential
Figure 4-12: Tambient throughout the year 2013.
Another effect of increasing temperature, although far less substantial, is the increase in Isc .
The band gap2 of the solar cell decreases with increasing temperature, resulting in a higher
absorption of photons and thereby an increase in the net current.
Ambient temperature
As described in the previous section, the temperature of a solar cell has an impact on its
performance. Therefore, it is important to know the cell temperature of the cells in a PV
module to make an accurate estimate of the power output of the module.
The first parameter that will influence the cell temperature is the ambient temperature. The
meteorological data from Royal Netherlands Meteorological Institute (KNMI) also provides
the ambient temperature in minute-resolution from the same weather station in Cabauw. This
ambient temperature fluctuates throughout the year, as indicated in Figure 4-12, influencing
the cell temperature and thereby the PV module performance.
Wind speed
Wind speed is the second parameter that will influence the cell temperature; it may cool the
PV cells through convective heat transfer. Therefore, also the data on the wind speed from
the weather station is considered. This data is measured at an height of 10 meters in intervals
of 10 minutes. The fluctuations in the wind speed in time are made visible in Figure 4-13,
where the wind speed is plotted for the first week of November 2013.
Since the anemometer used to measure the wind speed is located at higher height than the
PV module on the infotainment spot, the wind speed is scaled down by [31]:
1
Hmodule

5
Vwind,module = Vwind,anemo (4-26)
Hanemo
2
The band gap of a material corresponds to the minimum amount of energy that an incident photon should
have in order for it to be absorbed by the material.

4-2 PV output 57
Figure 4-13: Wind speed throughout the first week of November 2013.
where Hmodule and Hanemo are the module height and anemometer height respectively. The
exponential factor of 15 is determined by the landscape surrounding the PV module. For
city-like surroundings a factor of 31 is usually used, however the wind speed in the vicinity
of the high EEMCS building is considered similar to an open field - based on the velocity
ranges - resulting in the factor 51 . For Hmodule 0.70 meters is used, corresponding to half of
the infotainment spot height H.
Irradiance
The irradiance on a PV module not only directly influences its performance, as discussed
in section 4-2-2, but it has an indirect effect as well. When the PV cells absorb photons,
some energy of these photons will be lost in the solar cell as heat, leading to a higher cell
temperature. Irradiance is therefore the third parameter that influences the cell temperature.
Estimating cell temperature
Using the three parameters as discussed in the previous sections, an estimate is to be made
on the cell temperature and accordingly the PV module performance. Different models can
be found in literature to make this estimation of the cell temperature. In this section, three
models will be discussed: the Nominal Operating Cell Temperature (NOCT)-model, the
INOCT-model and a temperature model by Duffie & Beckman.
NOCT
NOCT is defined as the mean solar cell junction temperature reached by open circuited
cells, installed on an open rack, under the following standard reference environment (SRE)
conditions: (a) solar irradiance on the cell surface is 800 W m-2 , (b) ambient temperature is
20o C, and (c) the wind speed is 1 m s-1 [32]. The cell temperature Tcell is estimated from the
ambient temperature, the irradiance and the NOCT in the following way [28]:

58 PV Potential
Gcell (t)
Tcell,N OCT (t) = Tambient (t) + (N OCT − Tambient (N OCT )) (4-27)
Gcell (N OCT )
Gcell (t)
Tcell,N OCT (t) = Tambient (t) + (N OCT − 20) (4-28)
800
INOCT
The Installed Nominal Operating Cell Temperature (INOCT)-model makes a correction for
the fact that the PV module is mounted directly on the curved surface, so that the backside
of the PV module is not open to the wind (as is the case in the NOCT-model). The correction
for such direct mount is made by adding 18o C to the NOCT [31]:
Gcell (t)
Tcell,IN OCT (t) = Tambient (t) + (N OCT − 2) (4-29)
800
Duffie & Beckman
The actual wind speed at the location of the infotainment spot is not taken into account
by the INOCT-model. To make a correction for the actual instantaneous wind speed at the
PV module, Duffie & Beckman developed another correction factor, superimposed on the
INOCT-model. The equation for the cell temperature, including this correction factor, is
presented in Eq. (4-30) [33]:
Gcell (t) 9.5 ηc

Tcell,D&B (t) = Tambient (t) + (N OCT − 2) · 1− (4-30)
800 5.7 + 3.8 · Vwind,module τα
where:
ηc = the solar cell efficiency under STC,
τ α = the effective transmittance-absorptance product.
When τ α is multiplied by the incident radiation it yields the energy that is absorbed by the
cell. The value for τ α is not generally known, but an estimate of 0.9 can be used without
serious error since the term ηc /τ α is small compared to unity [33].
The relationship between the cell temperature and the wind speed, according to Eq. (4-30)
is plotted in Figure 4-14. The ambient temperature is fixed at 20o C and the irradiance on
the module at 800 W m-2 , corresponding to the INOCT-model. The INOCT is plotted in the
same figure for comparison. A solar cell with N OCT = 48o C and ηc = 0.1 is used for the
calculation, under the assumption that τ α = 0.9.
In Figure 4-15 the module temperature - the average of the segment temperatures - is es-
timated by the NOCT, INOCT and Duffie & Beckman temperature models for the day of
September 5th . All models are show a strong relation between the module temperature and
the irradiance on the module (refer Figure 4-9c). The INOCT model, which adds 18o C to
the NOCT, estimates the highest module temperature. According to the Duffie & Beckman
model, the wind speed cools the module down to approximately the module temperature as
predicted by the NOCT model.

4-3 PV module output 59
Figure 4-14: Tcell as function of wind speed, for Tambient = 20o C and Gcell = 800W m−2 .
Figure 4-15: Tmodule during September 5th 2013.
4-3 PV module output
As discussed in the previous sections, the irradiance and cell temperature have an impact
on the performance of a PV module. The model proposed by Duffie & Beckman is used to
calculate the cell temperature of the PV module on the infotainment spot. The influence
of the irradiance and the cell temperature on the output of the PV module is consequently
modeled by the following equation [34]:
Gi (t)
Pmpp,i (t) = Pmpp,i (ST C) · · [1 + κ (Ti (t) − 25)] (4-31)
Gi (ST C)
with:
Pmpp,module (ST C)
Pmpp,i (ST C) = (4-32)
n
where:
Pmpp,i (t) = the instantaneous maximum power output by segment i [W],
Pmpp,i (ST C) = the maximum power output by segment i under STC [W],
Pmpp,module (ST C) = the maximum power output by the module under STC [W],
Gi (t) = the instantaneous irradiance on segment i [W m-2 ],

60 PV Potential
Gi (ST C) = the irradiance under STC = 1000 W m-2 ,

κ = the temperature coefficient of the maximum power point (Pmpp ) [K-1 ],
Ti (t) = the instantaneous temperature of segment i [o C].
n = the number of segments.
The effect of the irradiance on the power output of segment i is expressed by the irradiance
fraction in Eq. (4-31), whereas the effect of temperature is visible in the factor with param-
eter κ. This temperature coefficient of Pmpp is a property of the PV module and differs per
solar cell technology. The value of κ is determined by the producer of the PV module and
mentioned in the datasheet.
The total power output by the PV module depends on the configuration (refer Figure 4-10).
For configuration A (Figure 4-10a), the cell that produces the least current will be limiting
for all cells connected in series. The power output by the module is then calculated by:
Pmpp,module (t) = n · Pmpp,i,min (t) (4-33)
For the configuration where the cells are all producing the same current (Figure 4-10b), the
power output by the module in that case is calculated by:
1=n
X
Pmpp,module (t) = Pmpp,i (t) (4-34)
i=1
4-4 Flexible PV modules
Although there is quite a number of flexible PV modules on the market, only a limited
selection is considered for implementation on the infotainment spot. The reason for this is
the size restriction for the PV module: it should not be longer than 2.1 m, as this module
length corresponds to the maximum footprint F of 1.40 m, with top and bottom offsets both
set to 0.05 m. The width of the module also has a limitation: it should be maximum 0.60
m. If the module would be wider than that, the infotainment spot would look too bulky and
lose the attractiveness in the design. A small number of different PV modules is selected and
they are discussed in the following sections.
4-4-1 Powerfoil® 115
The Powerfoil® 115 is a PV module produced by HyET Solar in the Netherlands. It consists
of 28 amorphous silicon (a-Si) solar cells connected in series. The module is 5.93 m long and
has a width of 0.325 m. HyET has confirmed however that for the case of the infotainment
spot this module could get a customized length. The rated power of the module is 116 W,
which results in a rated power of 39.1 W in the case of a 2.0 m long module.

4-4 Flexible PV modules 61
The cell configuration of the module is as shown in Figure 4-10b, which means no power is
lost due to differences in irradiance level between the segments. On the other hand, the fact
that this module has amorphous silicon solar cells, results in a relatively low rated efficiency
of 6.0%. The temperature coefficient of Pmpp , κ = −2.0 × 10−3 K-1 . The datasheet of this
module is attached in appendix A of this thesis.
4-4-2 Powerfoil® 165
The Powerfoil® 165 is produced by HyET Solar as well. It consists of 28 a-Si/microcrystalline

silicon (µc-Si) tandem solar cells connected in series. The module in the datasheet has the
same dimensions as the Powerfoil® 115. This module could also be customized in terms of
the length. The rated power of the module is 165 W, which results in a rated power of 55.6
W in the case of a 2.0 m long module.
It has the same cell configuration as the Powerfoil® 115, but the main difference is found in
the efficiency of the module. The double-junction of the solar cell results in higher absorption
of the sunlight, which in turn results in a higher rated efficiency of 8.6%. The temperature
coefficient of Pmpp of this module is slightly more negative, κ = −2.5 · 10−3 K-1 , meaning a
slightly larger effect of temperature on the PV module performance. For more details on the
parameters of this module, refer to the datasheet in appendix A of this thesis.
4-4-3 Powerflex™
This flexible PV module is produced by GlobalSolar, which is a company of Hanergy, China.

This module consists of 36 Copper Indium Gallium diSelenide (CIGS) solar cells connected in
series. The dimensions of the module are 2.017x0.494 m (lxw), just within the length restric-
tion. The module comes with different power ratings, but the one available on the market
in the Netherlands has a rated power of 90 W. This is substantially higher than the modules
produced by HyET, resulting mainly from the larger size and also from the slightly higher
rated efficiency of 9.0%. κ of this module is more negative however: −4.3 · 10−3 K-1 , meaning
a worse temperature effect on the power production by the module.
The module has a configuration as shown in Figure 4-10a. This means that the power that
the module produces is limited by the segment that produces the least power. Especially in
winter and summer, when direct light is irradiant on the module, this effect could be notable.
More characteristics of this module can be found in the datasheet in appendix A.
An exceptional feature is that each cell has its own bypass diode, to prevent hot-spots in the
module. Hot-spots are caused by partial shading of a PV module; if cells are connected in
series and one cell is (partially) shaded, the unshaded cells will reverse bias the shaded cell. If
this reverse bias is not limited, the extra power that all the unshaded cells are generating will
be dissipated in the shaded cell. This will heat up the shaded cell and may lead to melting
of solder and degradation of the solar cell [35]. A bypass diode creates an alternative current
path for a reverse biased solar cell, limiting the reverse bias voltage to 0.6-0.7 V (in case of a
silicon diode). In this way a limited amount of power will be dissipated in the shaded cell and
it will not create a hot-spot. The Powerflex™ module is thus exceptionally well protected
against possible hot-spots by having implemented a bypass diode for each individual solar
cell.

62 PV Potential
Figure 4-16: Emodule as function of footprint F .
Duffie & Beckman INOCT NOCT ηannual avg [%]

Powerfoil® 115 31.7 30.9 31.7 5.9
Powerfoil® 165 45.6 44.1 45.5 8.5
Powerflex™ 90 62.3 60.4 62.2 7.6
Table 4-1: DC yield prediction [kWh yr-1 ].
4-4-4 DC yield
The yearly DC yield, the energy that is produced by the PV module, Emodule , is calculated for
the different PV modules. For both the HyET modules, the yield is plotted as a function of the
footprint F in Figure 4-16. For comparison, the yield of the Powerflex™ module is indicated
in the same figure. The temperature model used for this plot is the Duffie & Beckman model.
To compare the three PV modules, in Table 4-1 the yield is listed for the footprint F = 1.4
m, as this is the footprint that is required by the Powerflex module. The predictions on yield
with different temperature models is included in the same table for comparison. Furthermore,
the annual average efficiency, calculated by Eq. (4-35), is included in the table as well.
Emodule
ηannual avg = · 100% (4-35)
Eaverage,global
where:
ηannual avg = the average annual efficiency,
Emodule = the DC yield produced by the PV module,
Eaverage,global = the irradiation averaged over the 4 segments, from Eq. (4-16).
The DC yield is that is predicted with the NOCT model is very close to the one predicted
with the Duffie & Beckman temperature model. This is in agreement with the observation
made in section 4-2-3, where the module temperature throughout the day is similar for both
models. When the module temperature is calculated by the INOCT model, the DC yield
prediction is 2.5%-3.3% lower due to the worse performance at higher module temperatures.

4-5 Conclusion 63
In general, the average annual efficiency is lower than the rated efficiency of the modules,
caused by the temperature effect on the PV module performance.
The Powerflex™ 90 obviously has the highest yield, due to its larger area compared to the
Powerfoil modules. The configuration of the cells in the Powerflex™ 90 causes the produced
power to be limited by the segment with the lowest irradiance. This causes the average annual
efficiency to be lower than the Powerfoil® 165 module, even though the rated efficiency is
higher. If the cells were configured in the same way as in the Powerfoil modules, a DC yield
of 72.0 kWh yr-1 could be reached, 15.6% more than in the actual cell configuration. Also,
the average annual efficiency in that case would be 8.8%, higher than that of the Powerfoil®
165 module.
4-5 Conclusion
In this chapter the curvature of the infotainment spot is modeled by a number of straight
segments. Four segments appeared to be enough to give an accurate representation of the
curved surface. The effect of irradiance and temperature on a PV module is discussed and
3 PV modules are introduced as possible candidates for the infotainment spot. Based on
different temperature models a prediction is made on the annual energy production for these
PV modules. The decision which PV module to use in the infotainment spot cannot be
decided only by this DC yield prediction, but depends on the energy needs of the system as
well. The process of determining this energy demand is discussed in the next chapter.

64 PV Potential

Chapter 5
Building the system
In the previous chapter, the annual amount of energy produced by a photovoltaic (PV)
module on the infotainment spot is computed. In order to decide which PV module would
suit the infotainment spot best in terms of energy supply, the total energy demand of the
system needs to be determined. This chapter describes the process of this calculation and the
required measurements that are performed in sections 5-1 and 5-2. In the next section, 5-3,
the modeling of the system is described to analyze the performance of the system throughout
the year. The final section of this chapter is dedicated to the fabrication of a prototype.
5-1 Topology
The electric components of the system can be arranged in various ways, in other words: the
system could have various topologies. Most of the components have to be present in each
system topology:
the PV module, the tablet, the usb ports, a battery as energy storage and a charge controller
to regulate the energy flows properly. The freedom that is left is mainly reflected in the power
converters and inverters that would be needed for different topologies:
1. The tablet could be connected to the associated Alternating Current (AC) charger. This
would mean that the Direct Current (DC) system would need an inverter to create an
AC voltage between 100V and 240V, the input voltage range of the AC adapter. The
USB ports could then be connected to the same AC ‘micro-grid’ with a general AC USB
charger, widely available on the market.
2. A second possible system configuration is similar to the previous one, but with the USB
ports connected to the DC system. These types of USB chargers are also available on
the market, especially for a DC system voltage of 12V or 24V.
3. The third option is to connect the USB ports and also the tablet to the DC system,
because the input power requirement of the tablet itself is DC power anyways.

66 Building the system
Input tablet Charge controller

Vmin 10.5 11.5
Vmax > 15.0 14.7
Table 5-1: DC voltage range of the tablet and a typical 12V charge controller.
Figure 5-1: Topology of the system.
In the first two system topology options, there would be the need for an inverter to convert
the DC system to an AC ‘micro-grid’. In this conversion process from DC to AC, there will be
some energy losses. To feed the tablet with power, the rectifier in the associated AC charger
converts the AC signal back into DC, with the accompanying conversion losses. This is why
the third topology would be the most favorable, as both these conversion losses would be
eliminated.
The DC input voltage of the tablet is rated at 12V, which is the rated voltage of a large
variety of batteries as well. This would mean that the third option could be realized in terms
of rated voltage level. The voltage of a battery varies however, depending on its state of
charge (SOC) and the operation, as will be discussed in section 5-3-1. To make sure that in-
put of the tablet is compatible with the voltage range of a typical 12V battery, measurements
have been performed. The tablet is connected to a DC power supply and the voltage level of
the power supply is varied over the operating voltage range of a typical 12V battery. A charge
controller in the system will make sure that the operating voltage will never become below
Vmin for deep-discharge protection. The highest voltage that a charge controller would apply
over the battery, Vmax , is during a stage of so-called equalization charge. Typical values are
presented in Table 5-1, together with the measurement results.
From Table 5-1 it is obvious that the possible charge controller voltage levels lay well within
the input voltage range for the tablet. This means that the tablet can be directly connected to
a 12V DC system without any problems. The system topology that is chosen is therefore the
third option, without an AC ‘micro-grid’, to eliminate extra conversion losses. A schematic
view of the topology is presented in Figure 5-1. To highlight that the USB charger converts
the DC system voltage from 12V to a USB voltage of 5.00V ± 0.25V, the converter within
the USB charger is drawn separate from the USB ports.

5-2 Load profile 67
Figure 5-2: Setup to measure power demand of the toughpad FZ-A1.
5-2 Load profile
The electrical components that are consuming power in the system, the loads, consist of the
tablet and the USB charger. The amount of power that these components consume depends
on several factors. The most important factors for the tablet are the brightness setting of
the screen and the fact if the tablet is actively used or not. The power that the USB charger
consumes depends on whether a device is being charged or not, the type of device and the state
of charge of its internal battery. To make an accurate estimate on the power consumption of
both electric loads (and therefrom also the energy consumption), these factors are discussed
in the following sections.
5-2-1 Power demand Toughpad FZ-A1
One of the most decisive factors in the power consumption of the tablet is the brightness
setting during operation. Four different possible settings that could be used are considered:
• The screen is always on at the highest possible brightness.

• The screen is always on, but the brightness is automatically set by the tablet.
• The screen switches off after 30 minutes not being used, and during use it has the
highest possible brightness.
The possibilities are listed in order of desirability, but this is at the same time the order of
highest energy consumption as well. The choice for a certain setting will be determined from
the simulation results from section 5-3. The power demand by the tablet is measured for
all of the above-mentioned settings. In the same measurement setup (see Figure 5-2), the
influence of active use of the tablet is measured as well. In the system topology, the tablet
is permanently connected to the charge controller, which will make sure the battery of the
tablet is always fully charged by trickle charging. That is why during the measurements,
the battery of the tablet was fully charged as well to measure the same demand as it would
have in the real system. Both voltage and current are measured with an ELV® 72816 digital
multimeter over a period of 120 minutes with 1 minute resolution. The average current and
voltage are multiplied to obtain the average power consumption in that brightness setting;
the results are presented in Table 5-2.
The results show that even when the screen is switched of, the tablet still consumes 1.7W
of power, which is 1.2W less than in the lowest brightness setting. In the highest bright-
ness setting, the power consumption is 5.6W and if the tablet is actively used, the power

Power [W]
screen off 1.7
lowest brightness 2.9
highest brightness 5.6
actively used +2.1
Table 5-2: Power consumption of the toughpad FZ-A1.
Figure 5-3: Power demand of the toughpad FZ-A1 in automatic brightness mode.
consumption rises on average with 2.1W. This additional power consumption is measured
in the highest brightness setting and assumed independent of the brightness setting. If the
automatic brightness setting of the tablet is to be used, the brightness profile of the tablet
during a day is determined. The orientation of the tablet in this measurement was as in the
proposed locations: the screen tilt angle (θscreen ) was set to 65o and the screen was facing
North. In Figure 5-3 the results of the brightness measurements are schematically presented.
The brightness of the screen is dependent on the amount of ambient light, measured by a
sensor on the front-side of the tablet. The period of ambient light on the tablet on a short
winter day would result in a different daily load profile than for a long summer day. The
approximation is therefore made that the brightness profile of the tablet is a function of the
sunrise and sunset time of the day, so that the load profile for automatic irradiance can be
determined for each day of the year. The second approximation made is that the brightness
profile is independent of the daily weather conditions, so solely a function of the sunrise and
sunset time. If in reality this is not always the case, a certain safety margin in the load profile
is built-in, as the measurements were performed on very sunny day (16/08/2014). As pre-
sented in Figure 5-3, the tablet works in the lowest brightness setting between 1 hour before
sunset up to 1 hour after sunrise. In a period of 75 minutes it will then increase the bright-
ness up to the maximum level, which it will maintain until 135 minutes before sunset. The
brightness level then goes down in the same time period of 75 minutes to the lowest brightness.
With the power consumption of the tablet in all situations known, the times that people will
use the tablet throughout the day is estimated. A distinction is made between weekdays and
weekend/holidays; during the weekend or holiday it is assumed that the tablet is used less
than on weekdays. Furthermore a distinction is made between ‘summer’ and ‘winter’ days; for

5-2 Load profile 69
Figure 5-4: Load profile of the tablet on a ’Summer Weekday’.
Figure 5-5: Setup to measure power demand of the USB ports.
months April-September the tablet is assumed to be used more often than in the remaining
months. The combination of these two distinctions results in 4 different daily usage profiles.
For details on the assumptions of the four different daily load profiles, see Table 5-4.
The load profile of the tablet is consequently created for the different brightness settings. In
Figure 5-4 the results is shown for a certain day in the year. The 75-minute period of rise and
fall in daily power consumption (in setting 2) is simulated by 75-minute periods of average
power consumption. To give an idea on the yearly power consumption a simple calculation
can be made: let’s assume the power consumption of the tablet is 5.6 W continuously (a
simplification of setting 1), this would mean an annual energy consumption of 49 kWh. This
energy consumption of the tablet alone would already exceed the yearly DC yield of both
Powerfoil® modules with the largest footprint (refer Table 4-1) without considering system
losses and seasonal variations in power production.
5-2-2 Power demand USB ports
With the load profile(s) of the tablet known, the power consumption of the USB charger is
to be determined. The measurement setup is shown in Figure 5-5. The power consumption
of the USB charger depends obviously on the fact whether a device is connected to it or not.
Furthermore, the type of device may influence the power consumption (Apple products that

Figure 5-6: Power consumption of the USB charger as function of the battery SOC.
PUSB [W]
no load 0.08
Samsung Galaxy S3 5.77
Apple iPhone 4S 5.87
Apple iPad Air 5.98
Table 5-3: Power consumption of the USB charger.
can be charged via USB cable generally need a 2.1A USB port). The SOC of the battery
of the device also influences the power consumption of the USB charger, because the charge
regulator within the device demands a lower current when the battery of the device is near
100% SOC. In the performed measurements, the influence of all these factors on the power
consumption of the USB charger is investigated. The power demand of the USB charger,
while a Samsung Galaxy S3 is charging, is plotted as function of the SOC of the phone
battery in Figure 5-6. It can clearly be seen that the power demand drops after a certain
SOC is reached, caused by the charge regulator in the phone. The maximum power demand
by the USB charger is presented in Table 5-3 as well, together with the maximum power
demand when other devices are connected.
It is remarkable that although the Apple products need a 2.1A USB port (they did not
charge in the 1.0A USB port), they both demand only slightly higher current than the Sam-
sung phone.
For creating the load profile, it is assumed that when a phone is connected to the system, that
the power demand of the USB charger equals 5.9 W. The times that a user has connected
his/her smartphone to the infotainment spot is assumed to be depending on the type of day
as defined previously in section 5-2-1 and have a correlation with the times that the tablet is
used. The assumptions of the USB charger usage are presented in Table 5-4 and for a certain
type of day the load profile of the USB charger is plotted in Figure 5-7.

5-2 Load profile 71
Figure 5-7: Load profile of the USB charger on a ’Summer Weekday’.
Summer Summer Winter Winter

Weekday Holiday Weekday Holiday
Tablet 08:20-08:30 12:00-12:10 08:20-08:30 12:00-12:10
12:30-12:50 14:00-14:10 12:40-12:50 14:00-14:10
13:10-13:20 16:00-16:10 13:10-13:20
14:10-14:20 17:30-17:40
15:30-15:40 17:50-18:00
17:30-17:40
17:50-18:00
USB 13:00-13:10 14:00-14:10 15:30-15:40 16:00-16:10
15:30-15:40 16:00-16:10 17:30-17:40
17:30-17:45
Total tablet [min] 80 30 50 20
Total USB [min] 35 20 20 10
Table 5-4: Assumed usage of the infotainment spot.

Figure 5-8: Total daily load profile for a ’Summer Weekday’.
Setting 1 Setting 2 Setting 3

Etablet 49.6 34.9 19.3
EUSB 1.5 1.5 1.5
Eload 51.1 36.4 20.8
Table 5-5: Annual energy demand by the loads for different screen brightness levels [kWh yr-1 ].
The load that is seen by the charge controller consists of the power demand of both the tablet
and the USB charger. For a ‘Summer Weekday’ the total load profile is plotted in Figure 5-8.
With the load profiles for the three different brightness settings known, the total annual
energy demand is computed and shown in Table 5-5. Obviously, the brightness setting of
the tablet does not influence the annual energy demand by the USB ports, which equals
1.5 kWh. When the tablet operates in the highest brightness throughout the year, it would
add up to a total load demand of 51.1 kWh yr-1 , more than either of the Powerfoil® could
provide annually, even without all other system losses excluded. When the tablet operated
in ‘automatic brightness’ mode, it would result in a lower annual load demand of 36.4 kWh.
It would be preferable to choose this option over the third one, where the screen switches off
after 30 minutes of not being used. This would lead to an even lower energy demand of 20.8
kWh annually, but it would also lower the attraction for people passing the infotainment spot
to use it.
5-3 System modeling
In order to take system losses into account and to analyze the performance of the system
throughout the year, a model of the system is made in Simulink. Simulink is a graphical
programming tool developed by MathWorks, for modeling, simulating and analyzing dynamic
systems. The model will give an indication of the required battery size to make sure that the
system is provided with sufficient energy to operate continuously throughout the year.

5-3 System modeling 73
5-3-1 Storage
The main reason that the system modeling is performed in Simulink, is that a battery model
is available within the Simulink library. This battery model implements a generic dynamic
model parameterized to represent most popular types of rechargeable batteries. The param-
eters of the equivalent circuit can be modified to represent a particular battery type, based
on its charge and discharge characteristics [36].
The most common type of battery found in stand-alone solar systems comprises rechargeable
lead-acid-batteries. These are the most cost-effective and can handle large and small charging
currents with high efficiency [37]. As the tablet requires a DC input voltage of 12 V and be-
cause it is directly connected to the charge controller, the battery needs to have a rated voltage
of 12 V. The parameters in the Simulink battery model are thus set to a 12V lead-acid-battery.
The internal resistance of the Simulink battery model is modeled as a constant, while in
reality it is a function of the SOC of the battery. This simplification however would only
be problematic if the purpose of the model is to get insight in the internal processes of the
battery. On system-level, it has a negligible effect on the total battery performance and is
therefore not of any interest.
Another simplification of the Simulink model is that the SOC of the battery is independent
of the temperature. In reality, however, the capacity - and therefore the actual SOC - of
a battery decreases for lower temperatures. This effect is noticed at system-level and is
therefore implemented in the model. The assumption is made that the battery temperature
is equal to the ambient temperature. Consequently, the battery SOC is multiplied with a
temperature coefficient to obtain the actual SOC of the battery at that particular moment.
The temperature coefficient is determined from Figure 5-9, the temperature dependence of
the capacity of a typical lead-acid battery [38].
SOCactual (t) = SOCsimulink (t) · T C (5-1)

where:
T C = the temperature coefficient of battery capacity.
5-3-2 Charge controller
The charge controller in a stand-alone PV system has a variety of functions. It controls

the power flowing towards or from the battery depending on the power provided by the PV
module and the power demand by the load. Furthermore, it makes sure that the battery
is not overcharged or too deep discharged and it provides information on the SOC of the
battery. The charge controller that is chosen for the infotainment spot is the Steca Solarix
MPPT 1010. The choice was mainly determined by two properties:
1. It has a Maximum Power Point Tracking (MPPT) mechanism. An algorithm is imple-

mented to search from time to time for what operating voltage the PV module power

Figure 5-9: The effect of temperature on battery capacity, adapted from: [38].
Figure 5-10: Setup to measure efficiency of the charge controller.
is maximum, in other words: it searches for the maximum power point voltage (Vmpp ).
In this way, the operating voltage is not determined by the battery voltage as in charge
controllers without MPPT.
2. The efficiency for low load power is relatively high in comparison with other MPPT
charge controllers.
Furthermore, it has quite a number of features and electronic protection functions as temper-
ature compensation, short circuit protection and reverse current protection. A full list of its
features can be found in the datasheet in appendix A. The charge controller itself consumes
a small amount of power as well and therefore measurements are performed account for these
losses later in the Simulink model. The setup to determine the efficiency of the charge con-
troller is shown in Figure 5-10.
By varying the DC load, the efficiency of the charge controller is determined for two situations:
during charging and during discharging of the battery. The equation for the charge controller
efficiency in those both situations are given in Eq. (5-2) and Eq. (5-3) respectively.

Figure 5-11: Charge controller efficiency as function of the load power.
Pbattery + Pload
ηcc,charging = · 100% (5-2)
Pmodule
Pload
ηcc,discharging = · 100% (5-3)
|Pbattery | + Pmodule
The definitions of the different power flows and their directions is shown in Figure 5-12;
Pmodule is the power from the PV module towards the charge controller; Pbattery is the power
from the charge controller towards the battery; Pload is the power from the charge controller
towards the load(s). Pmodule and Pload can only have values larger than or equal to zero, while
Pbattery can have both positive and negative values, representing charging and discharging of
the battery respectively. In the Simulink model of the charge controller the same definitions
will be used for the sake of continuity.
The results of the efficiency measurements are presented in Figure 5-11 as function of the
load power. The dashed lines give the linear approximations that are used in the Simulink
model of the charge controller.
With the efficiency known, the charge controller is modeled in Simulink. The inputs for the
module are Pmodule and Pload and the outputs are Pbattery , Pdump and Pdef icit . Both the dump
and the deficit are only implemented in the model to analyze the system performance. In
reality no dump is needed: if the PV would produces more power than the charge controller
can send to the battery and the load, the charge controller makes sure that the PV module
operates in a lower power point. The deficit is implemented to analyze how much power is
not provided to the load at times that there is demand.

Figure 5-12: Definition of the energy flows in/from the charge controller.
Figure 5-13: Control strategy for the charge controller.
The following control strategy is implemented in the charge controller model and is schemat-
ically shown in Figure 5-13. The SOC refers to the state of charge of the battery in the
system, modeled by the Simulink battery model. The matlab code for the charge controller
is attached in appendix A.
• When Pmodule > Pload the battery is charged by the excess power if the SOC is 80% or
less.
• When Pmodule > Pload and the SOC is higher than 80%, the excess power will be
dumped.
• When Pmodule ≤ Pload the battery is discharged with the shortage in power if the SOC
is at least 20%.
• When Pmodule ≤ Pload and the SOC is lower than 20%, the charge controller disconnects
the load, so that the battery won’t be discharged further. The power not provided to
the load in this situation is called Pdef icit .

Battery capacity [Ah]

Brightness setting 1 1500
Table 5-6: Estimate of battery size depending on screen brightness setting.
First, the charge controller compares the load demand with the supply from the PV module.
Based on the outcome, it would either need power from the battery - when Pmodule ≤ Pload
- or send power into the battery - when Pmodule > Pload -. An extra restriction, however,
decides whether the action will be performed or not, based on the SOC of the systems bat-
tery. To protect the battery from deep discharge and thereby prolonging the lifetime of the
battery [38], the charge controller does not discharge the battery further than 20% SOC; in
that case, the load will be disconnected from the circuit and the power not provided to the
load is called Pdef icit . In reality, the Steca Solarix MPPT 1010 will disconnect the load from
the circuit when the battery voltage comes below 11.5 V1 . It will reconnect the load to the
circuit when the battery voltage reaches 12.5 V again2 . The model does not take into account
the difference between the LVD and LVR, but in the charge controller in reality both voltage
levels could be set to the same value.
To prevent it from overcharging and thereby again prolonging the lifetime [38], it also does
not charge the battery further than 80% SOC. In reality, the Steca Solarix MPPT 1010 will
stop charging the battery when its voltage reaches 13.9 V. The value of this end-of-charge
voltage can also be adjusted.
The MPPT function of the charge controller will search for the Vmpp . In order to do so, from
time to time it will operate briefly in different voltage levels to find the optimum. During
these periods it would operate at lower power levels than the maximum power point (Pmpp ).
To account for these losses and to account for the fact that it might not always operate exactly
at the Pmpp , an average MPPT efficiency of 95% is implemented in the model.
5-3-3 Results
The simulation is performed for different brightness settings of the tablet. The main param-
eter that has been reviewed is the SOC of the battery, to get an idea of the required battery
size. The size of the battery is varied in the simulation such that the SOC did not go below
20%. In Figure 5-14 a typical SOC variation throughout the year is shown. Table 5-6 shows
the results for the different brightness profiles in combination with the Powerflex™module.
Due to the large seasonal variations in irradiance, the battery storage mainly functions as
seasonal storage. This means that during a large part of the year the battery is fully charged
(to 80% SOC) and during winter the SOC drops significantly. From Table 5-6 it is obvious
1
This is the Low Voltage Disconnect (LVD) voltage. Its value can be adjusted with the Steca PA RC100
Remote control
2
This is the Low Voltage Reconnect (LVR) voltage and can be adjusted as well.

Figure 5-14: Typical variations in SOC of the battery throughout the year.
Battery capacity [Ah]

Powerfoil® 115 500
Powerfoil® 165 300
Powerflex™ 90 150
Table 5-7: Estimate of battery size depending on the PV module.
that the third brightness setting needs to be implemented to have a reasonable battery size.
This setting switches the screen off after 30 minutes of non usage.
The simulation is also run with the other two PV modules in their maximum footprint con-
figuration (F = 1.40 m). Again the battery size is varied such that the SOC would not reach
20% or less. The results are presented in Table 5-7.
Although the DC yield of the Powerflex™ 90 is only 37% larger than that of the Powerfoil®
165, the required battery size of the latter is approximately twice as big. To keep the battery
size of the infotainment spot as small as possible - because of costs and space restrictions -
the Powerflex™ 90 is the best option out of the three PV modules.
5-3-4 System performance
The brightness setting of the tablet is set to the third option: the screen switches off after 30
minutes not being used, and during use it has the highest possible brightness. Furthermore,
the Powerflex™ 90 module is chosen out of the three PV modules. The system performance
is now evaluated for different sizes of the battery storage.
One way to quantify the reliability of supply is by a parameter known as the loss-of-load-
probability (LLP), defined as the ratio between the estimated energy deficit and the energy
demand over the total operation time of the installation [28], see Eq. (5-4).

Battery size [Ah] LLP [%] PUE [%]

80 1.29 56.1
100 0.52 55.8
120 0 55.7
150 0 55.7
Table 5-8: System performance as a function of battery size.
Edef icit
LLP = · 100% (5-4)
Eload
where: Z
Edef icit = Pdef icit (t)dt (5-5)
Z
Eload = Pload (t)dt (5-6)
With the LLP calculated for different battery sizes, the system sizing can be performed based
on the reliability of the supply. At the same time, for each battery size, the percentage of
unused energy (PUE) is calculated by Eq. (5-7).
Edump
PUE = · 100% (5-7)
Emodule
where: Z
Edump = Pdump (t)dt (5-8)
Z
Emodule = Pmodule (t)dt (5-9)
In the simulations the SOC at the start of the year (t=0 s) was set such that it was equal to
that of the end of the year (t=31536000 s). The results of the simulations are presented in
Table 5-8.
It should be noted that the for the simulations only the data from the year 2013 is used as
input. The annual irradiation might be similar between years, as concluded earlier, but the
variation in the annual irradiance profile could result in different simulation results. These
variations between different years are assumed small and are therefore neglected.
Another point of interest is that in the simulations, the battery of the tablet itself is not
taken into account. In reality, this battery is always fully-charged, as the tablet is perma-
nently connected to the charge controller, which will trickle charge the tablet battery. At
times that the charge controller disconnects the load from the circuit, the tablet will still be
able to work for some hours, which would decrease the LLP. However, at the moment that
the charge controller reconnects the load to the circuit, the load will be higher as it wants to
recharge the battery of the tablet. Therefore, the influence of the tablets battery on the LLP
is neglected.

Figure 5-15: The SOC in January for an 80 Ah battery.
From the table it can be seen that for this specific year a battery of 120 Ah would be sufficient
to have zero loss-of-load. Oversizing the battery has no influence on the simulation results,
for a battery of 150 Ah even the PUE is the same. An LLP of 1.29%, in the case of an 80
Ah battery, might not seem like a lot, but the simulation shows that in January the SOC
of the battery stays at ±20% for about 6 days, see Figure 5-15. This would mean that the
infotainment spot would not operate for about 6 days (minus the hours that it works on its
own battery), which would not be acceptable for the infotainment spot.
Discussion
When sizing the system it is necessary to evaluate the performance of the components in the
longterm. The lifetime of a battery is generally expressed in numbers of cycles, or cycle life.
Cycle lifetime is defined as the number of charging and discharging cycles until the actual
remaining battery capacity drops below 80% of the nominal capacity [38]. If one would want
the battery to have a certain capacity after its cycle life, the original capacity should therefore
be 25% larger. This factor of 1.25 is referred to as aging factor.
The number of cycles of a battery depends mainly on two factors: the depth-of-discharge
(DOD) and the temperature. The influence of these two factors on a typical lead-acid battery
are shown in Figure 5-16 [38].
For the case of the infotainment spot it is quite hard to estimate the number of cycles. The
number of cycles in a year can be approximated by one annual cycle with 60% DOD, with 365
average daily cycles with a very small DOD superimposed on the annual cycle. But even in
the case of a daily cycle of 10% DOD - an overestimation of the actual daily DOD - it would
result in over 4000 daily cycles, extrapolating Figure 5-16 for the highest battery temperature
of 45o C. The true cycle life of the battery is hard to estimate.

Figure 5-16: The influence of DOD and temperature on battery cycle life [38].
Not only the battery performance decreases over the years, the output of the PV module
dwindles over the years as well. The Powerflex™ 90 guarantees a minimum nominal power of
90% of the original value after 10 years and 80% after 25 years, as indicated in the datasheet
in appendix A.
A rough estimate of the required battery capacity in reality can be made by assuming that
the power decrease of the PV module is linearly related to the battery capacity, and if the
battery cycle life is assumed to be 10 years3 :
1
Crequired = Csimulation · 1.25 · (5-10)
0.9
where:
Csimulation = the preferred capacity from the Simulink simulation,
1.25 = the aging factor of the battery for its end of cycle life,
1
0.9 = the aging factor of the PV module after 10 years.
For the case of the simulation with a 120 Ah battery, this would result in a battery size of
approximately 167 Ah. More accurate calculations can be made when more information on
the cycle life of the battery is known. Furthermore, the simulations can be made more accu-
rate by altering the input Pmodule : the instantaneous maximum power output by segment i,
Pmpp,i (t) in Eq. (4-31), could be multiplied with the aging factor of the PV module.
Battery technology
The most common battery type used in solar energy systems is the lead-acid battery with grid
plates and fluid electrolyte. Batteries with fluid electrolyte are manufactured in an enclosed
design as an inseparable unit with housing. Only the sealing plugs can be opened to check
and top up the electrolyte level. In contrast, gel batteries do not require any sealing plugs
since under normal operating conditions there is no need to add water. As a result, they are
built as enclosed batteries and are fitted with safety valves, which allow any gas to escape
that results from overcharging. Gel batteries are therefore maintenance free [37].
3
This cycle life is merely an example to make the calculation possible.

Figure 5-17: Battery size and capacity, data from [39].
The infotainment spot is designed to operate autonomously in every sense of the word, so
regular maintenance is undesirable. Therefore the recommended battery type is a Valve Reg-
ulated Lead-Acid (VRLA) battery. Gel batteries have a longer service life, but are more
expensive than lead-acid batteries with fluid electrolyte. The application field for these bat-
teries clearly opens in the direction of all-year-round use with several years of service life [37].
To get an idea of the size of such VRLA battery, Figure 5-17 shows the size of typical VRLA
batteries [39] with respect to their capacity.
5-4 Building the infotainment spot
This final section of the thesis report describes the fabrication of the infotainment spot. A full-
scale prototype of the design is made, including the operating electronic system components.
The prototype gives a clear view on the basic structure of the design as well as an opportunity
for people to get the user-experience. All the electronic system components are incorporated
and functional.
The components
The system is built up from components already available on the market, some with small
modifications. An overview of the components is given in Figure 5-18. The components in
Figure 5-18a - Figure 5-18f are available on the market and not modified in any way for the
implementation in the infotainment spot. The AC adapter provided with the tablet is used to
create the tablet charging cord, Figure 5-18g. The cord was cut at the DC side of the rectifier
and banana plugs are connected to both of the DC cables. Banana plugs are connected to the
PV module connectors and the cigarette socket as well, see Figure 5-18h and Figure 5-18i.

5-4 Building the infotainment spot 83
(a) Charge controller [40] (b) PV Module [41] (c) Tablet [16]
(d) Battery (e) USB charger [42] (f) Battery cables
(i) Cigarette socket

(g) Tablet charging cord (h) PV module connectors
Figure 5-18: The components used in the prototype.

Price
Panasonic Toughpad FZ-A1 €1028.40
Steca Solarix MPPT 1010 Charge Controller €197.95
Powerflex-90 CIGS Module €332.75
USB charger €9.85
150 Ah VRLA battery [43] €344.00
Total €1912.95
Table 5-9: Indication of the cost-price of the system components.
Cost indication
With the final material and production techniques still unknown, merely an indication of
the component costs can be given at this stage. The components and their costs are listed
in Table 5-9. The cables and connectors are not listed in the table, as they were readily
available in the PV Laboratory of the Photovoltaic Materials and Devices (PVMD) group of
Delft University of Technology (TU Delft). If those costs will be included, the component
costs would add up to approximately €2000.
The battery mentioned in the table is just an indication of the price level of a typical VRLA
battery. The actual battery is still to be purchased, and its size can be different based on the
final location of the infotainment spot. The charge controller would need to be suited to the
specific properties of the battery as gel batteries are highly sensitive to overcharging. The
charge cut-off voltage must be strictly adhered to so that gassing does not occur, which dries
out the battery [37].
5-4-1 The prototype
The prototype is build up from a wooden skeleton to which 10 mm thick hardboard plates
are attached. This construction is visible in Figure 5-19. The curvature on the front-side
and back-side are made from 4 mm thin semi-flexible hardboard. This material is also used
to build the encapsulation of the tablet to make sure that the screen would not reside very
far below the surface of the encapsulation. The USB charger is located at the side of the
infotainment spot, at approximately the same height as the touchscreen. The prototype is
presented in Figure 5-20.
The charge controller is placed behind the touchscreen - as visible in Figure 5-19 - so that
long wires would not be needed to connect it to both DC loads. For the same reason, the
PV module connectors are located at the top, where the cables to the charge controller are
entering the embodiment. The bottom area inside the embodiment is therefore completely
available for the battery storage. In case of the prototype a 26 Ah battery is used that was
readily available in the PV Laboratory of the PVMD group of TU Delft. This is not the
capacity needed for the actual product, but it functions as an energy buffer just to make sure
the prototype is operational.

5-4 Building the infotainment spot 85
Figure 5-19: The skeleton construction of the prototype.
Certain details are not yet implemented in the prototype; a lid would be wanted in front of
the USB ports to prevent rain and dust to enter the USB ports, clogging the openings and/or
creating a chance of a short circuiting in the DC USB adapter. For similar reasons, the bor-
der between the tablet and its encapsulation should be closed with the appropriate silicone
sealant to make it waterproof. Furthermore, the edges of the PV module and its junction box
should be covered to make sure nobody could easily harm or steal the PV module. As the
prototype functions as a demonstrator, it will probably not have a permanent, fixed location.
In order to move it around easily, the entire system is placed on wheels, which changes the
height of the screen slightly from the ideal case.
It is recommended that for the production of the final infotainment spot experts in production-
techniques, e.g. in the Electronic and Mechanical Support Division (DEMO) of TU Delft
will be consulted. They would be able to give suggestions on the implementation of above-
mentioned details as well as ideas on the best material and production techniques.

Figure 5-20: The prototype of the infotainment spot.

Chapter 6
Conclusions and Recommendations
In this thesis a solar powered infotainment spot was designed and optimized for implemen-
tation on the campus of TU Delft. The embodiment of the infotainment spot is designed so
that the flexibility of a PV module is visible for the user. The location of the infotainment
spot has been thoroughly analyzed in terms of shading to accurately estimate the irradiance
on the PV module. With the use of a PV model the output of the PV module is computed
throughout the year. Measurements have been performed to create a load profile of the sys-
tem and the operation of the system is simulated for several battery capacities.
The questions that formed the basis of this thesis work are presented here again and conse-
quently each of them is addressed:
1. How much of solar energy can be harnessed by an existing flexible photovoltaic (PV)
module based in the Netherlands?
(a) How much of solar energy is available throughout the year in the Netherlands?
(b) In what way can the amount of solar energy incident on a curved plane be accu-
rately estimated?
(c) How accurate is the estimation of diffuse irradiance with the general models found
in literature?
(d) How much of electrical power can be generated by a PV module throughout the
year?
2. Is it possible to rely entirely on the solar power throughout the year to meet the daily
load demand?
(a) How does an autonomous PV system fare in terms of meeting the energy demand
of the touchscreen of the infotainment spot throughout the year?
(b) To what extent is it possible to include additional electronic loads apart from the
touchscreen?

88 Conclusions and Recommendations
6-1 Conclusions
1.(a) How much of solar energy is available throughout the year in the Netherlands?
There is an average annual irradiation of 1.05 MWh m-2 on a flat horizontal surface in
the Netherlands, of which about 57% is originating from diffuse irradiance. If the surface is
tilted with an optimal angle of 31o , the maximum irradiance increases to 1.16 MWh m-2 ,
mainly due to a large increase in the direct component. When also the surroundings (of the
infotainment spot in particular) are taken into account, the global irradiation on a horizontal
plane decreases with 23.3% while for an optimally tilted plane this decrease is 24.8% at a
tilt angle of 33o .
1.(b) In what way can the amount of solar energy incident on a curved plane be accurately
estimated?
To make an accurate estimate of the solar irradiation on a curved surface, an approach of
segmentation is used: the curved surface is divided into straight segments, each with their
own tilt angle. The irradiance on 1 /4 of a circle is accurately described by 4 segments:
the annual irradiation calculated with 4 segments deviates 0.23% of the annual irradiation
calculated with 104 segments.
1.(c) How accurate is the estimation of diffuse irradiance with the general models found in
literature?
The diffuse sky models in literature generally assume a SV F that is solely a function of the
tilt angle of the surface. As concluded earlier, in the Netherlands the diffuse component is a
big fraction of the global irradiation on a surface and the estimation of this component there-
fore should be as accurate as possible. The SV F is therefore determined by 3D modeling of
the surface and its surroundings. The difference between the two approaches is caused by the
fraction of the sky that is blocked by the surroundings of the surface. For a horizontal plane
the SV F is determined to be less than 0.7, whereas the models in literature assume 1.0. The
deviation is a function of the surface tilt angle and ranges between 0.2 and 0.3.
1.(d) How much of electrical power can be generated by a PV module throughout the year?
The three different PV modules discussed are based on different solar cell technologies, with
different efficiencies and one of them has a different size than the other two. The power output
therefore differs from module to module. The power output is computed for each module by
taking varying irradiance and temperature effects into account. The average annual efficiency
of the modules drops slightly with respect to their rated efficiencies, resulting in the following
DC yields: 32 kWh yr-1 for Powerfoil® 115; 46 kWh yr-1 for Powerfoil® 165; 62 kWh
yr-1 for Powerflex™ 90. Also the cell configuration of the module has an impact on the PV
output. If the cells of the Powerflex™ 90 module were configured in the same way as in the
Powerfoil® modules, the yield would be 72 kWh yr-1 . This difference is caused by the dif-
ferent irradiance levels on different parts of the module, due to the shape of the infotainment
spot.

6-1 Conclusions 89
1. How much of solar energy can be harnessed by an existing flexible PV module based in the
Netherlands?
The amount of annual energy output by the PV module is only calculated for the specific case
of the infotainment spot. In this case, the highest yield was achieved by the Powerflex™
90 module: 62 kWh yr-1 . This is a fraction of 0.076 from the annual solar irradiation on
the module. Relatively, the Powerfoil® 165 performs better, with an annual yield of 71 kWh
yr-1 m-2 , compared to 62 kWh yr-1 m-2 of the Powerflex™ 90. The annual yield of the
PV modules could be increased by placing them on a flat surface under the optimal tilt angle
of 33o .
2.(a) How does an autonomous PV system fare in terms of meeting the energy demand of
the touchscreen of the infotainment spot throughout the year?
The energy demand of the touchscreen of infotainment spot is estimated based on a load
profile backed up by measurements of the actual components. Different brightness settings
of the touchscreen have been assessed, but only one of them resulted in a well-functioning
system with a reasonable battery size. The screen will be switched off automatically after 30
minutes of not being used; in this way the total annual demand by the tablet is estimated
to be 19.3 kWh yr-1 .
2.(b) To what extent is it possible to include additional electronic loads apart from the touch-
screen?
It is possible to implement other electronic loads apart from the touchscreen. In this case, a
USB charger is included in the infotainment spot, so the user can charge his/her smartphone
or tablet with clean solar energy via the usb connection. The measured power demand of
the USB charger in combination with an assumed load profile results in an annual demand
of 1.5 kWh yr-1 On one hand, more electronic loads could be included, because the annual
DC yield of the PV module is 62 kWh yr-1 , about 3 times as much as the loads demand
(system losses neglected). On the other hand, if the demand of an additional load would
be several kWh yr-1 , the battery size required for a full-time running system would increase
drastically. For example: another brightness setting considered would have 15.6 kWh yr-1
more demand, which would lead to a required battery of about 3 times larger capacity.
2. Is it possible to rely entirely on the solar power throughout the year to meet the daily load
demand?
Yes it is certainly possible to rely entirely on the solar power throughout the year to meet the
daily load demand. Even in the simulation where the LLP equals zero, the required battery
size in the simulation would be 120 Ah. If with a first order estimate the cycle life of the
battery and the performance of the PV module are considered, this would result in a required
battery size of 167 Ah. This battery would still easily fit inside the infotainment spot. When
additional loads are considered, the required battery size could however rapidly increase up to
a point where: either it is economically not attractive anymore to implement the additional
load, or the actual size of the battery would be too large to fit within the embodiment of the
infotainment spot.

6-2 Recommendations
The aim of this thesis report is to present a design of the infotainment spot that would be
able to power a touchscreen throughout the year as autonomous solar energy system. Part of
the purpose of the infotainment spot is to create awareness of PV (education) on the campus
of TU Delft.
6-2-1 Location
For this thesis project only locations close to the main entrance of EEMCS faculty, the faculty
where the PVMD research group is located, are considered. The irradiance on both proposed
locations however is severely lowered by the CEG and EEMCS buildings: they block a fraction
of the sky, which causes a lower diffuse irradiance on the module and depending on the position
of the sun it sometimes blocks the direct component of the sun as well.
Furthermore, both proposed locations will be mainly passed by students and employees of
the TU Delft. If a location like the entrance of the TU Delft Aula or Library would be con-
sidered, more visitors of the TU Delft would get a chance to experience the infotainment
spot. For these locations there may also be a higher yearly irradiation as the shading (and
blocking of the sky) by surroundings is likely to be less, but this would have to be investigated.
6-2-2 Design & Production
In this thesis project a specific shape is chosen as basis for the design of the infotainment spot.
Some of the dimensions are chosen for ergonomic reasons, others are based on the required
space for the PV module. A general idea on the materials to be used and the construction of
the infotainment spot are given in section 5-4. With the help of the prototype, the final design
could be optimized in terms of materials and fabrication techniques by calling in experts in
these areas.
6-2-3 Information
As the main focus of this thesis was on the feasibility of the infotainment spot, not much
research is done in the kind of information that will be provided by it. For now, a html/php
based webpage is created and implemented in the prototype, which is very easy to adapt.
I would recommend that for a specific location a research would be done to:
• the kind of users that would want to use the infotainment spot;
• the kind of information that each target group would like to get from the infotainment
spot;
• the way at which each target group wants to receive the information from the infotain-
ment spot.

6-2 Recommendations 91
This research might be done with corporation of students at the Industrial Design Engineer-
ing (IDE) faculty of TU Delft. To realize some of the results from this target group research,
the html/php webpage should be adjusted. One of the things that could be changed for sure,
is to make the webpage more interactive and less static.
Another outcome of this research would be a better underpinned load profile. The load profile
in this thesis report is created based on assumptions and estimates on the usage rather than
on actual research. One could never foresee the actual load profile of the infotainment spot,
but this research may result in a more accurate one. The influence of the load profile on the
annual energy demand however is small: 14.9 kWh yr-1 is demanded by the tablet when
the screen is switched off, which is over 71% of the total annual energy demand, including
the USB charger.
Furthermore, the entertainment part of ‘infotainment’ is not considered in this thesis report.
As mentioned in the first chapter, one could think of a speaker that can be connected via
bluetooth to a user’s smartphone or tablet. A variety of games can be included as well;
not only on the touchscreen itself, but one could think of a social game on campus, where
the infotainment spot could function as game-master. When one would think bigger, more
infotainment spots could be located throughout the TU Delft campus and connected together
in some way.
6-2-4 Additional loads
As might come forward from the user-experience research, some additional loads would be
desired, to make the user experience better and/or more interactive. Example giving some
LED strips could be implemented to make the infotainment spot more attractive to possible
users at nighttime. One should be careful in implementing these additional loads as they
may have a big impact on the required battery size. In that scenario, several options can be
chosen:
• The battery size stays the same, but the loss-of-load-probability (LLP) increases. In
this scenario it is likely that for some days in January or February the infotainment spot
cannot be used. The higher the load demand, the longer will the infotainment spot be
out of order.
• The battery size should increase to keep the LLP the same. This would mean that the
infotainment spot operates the same as without the additional loads. The downside of
this scenario, however, is that a larger battery comes at a higher price and there is also
a limit on the actual battery size in order for it to fit inside the infotainment spot.
• The power consumption of the additional loads can be limited by a controller in the
system. This controller could disconnect the additional loads and may also disconnect
the USB charger at certain times. This could be done based on the state of charge
(SOC) of the battery, e.g. below 40% SOC only the tablet would stay connected to
the circuit. The control could also be based on a timer: it could make sure that LED
strips for example are only switched on in the first 2 hours after sunset and 2 hours

before sunrise. The downside of this option is that some controlled circuitry should be
developed especially for the infotainment spot, but this doesn’t have to be expensive
and only has to be done once.
6-2-5 Battery
For the battery storage a Valve Regulated Lead-Acid (VRLA) battery with either gelled or
absorbed electrolyte is recommended, as it is a relatively cheap, maintenance-free battery.
The cycle life of the battery, however, is hard to estimate as it depends quite strong on the
depth-of-discharge (DOD) of the cycles and the battery temperature. Therefore, it would be
recommended that once the infotainment spot is realized, the battery parameters are checked
on a regular basis (e.g. annually or twice a year) to make sure that the battery still performs
as expected.

Appendix A
Appendices
Datasheet Panasonic Toughpad FZ-A1
Datasheet PowerFoil® 115
Datasheet PowerFoil® 165
Datasheet PowerFLEX™
Datasheet Steca Solarix MPPT 1010
The charge controller in Matlab

Durability n MIL-STD-810G certified (4' drop, shock, vibration, rain, dust, sand, altitude, freeze/thaw,
high/low temperature, temperature shock, humidity, explosive atmosphere)
n IP65 certified sealed all-weather design
n Magnesium alloy chassis encased with ABS and elastomer corner guards
n Port covers
n Raised bezel for LCD impact protection
n Pre-installed replaceable screen film for LCD protection
software n Android™ 4.0 (Ice Cream Sandwich)

n Calendar, Clock, Gallery, QuickSearchBox, Google Play™, Gmail™, Calendar Sync, Contacts
Sync, Google Maps™, Google+™, Google Talk™, Google Voice Search™, YouTube™, AuthenTec™
QuickSec, Over-the-Air Update Client, McAfee Security (Trial), Adobe Reader
warranty
n 3-year limited warranty, parts and labor
n Panasonic Utilities (Dashboard, User Button Manager, Device Management, Pen
Calibration, Hardware Diagnostics, Online Manual) dimensions & Weight6
10.5"(L) x 8.4"(W) x 0.7"(H)
LANGUAGES
n
n Android OS supports 58 language and locale combinations. Refer to
n 2.1 lbs.
http://developer.android.com/sdk/android-4.0.html#locs
Integrated Options7
CPU n Marvell® Armada PXA2128 1.2GHz Dual-Core Processor
n 4G LTE mobile broadband with satellite GPS
– Embedded hardware security processor
– FIPS 140-2 (Level 2) certified 1
ACCESSORIES7
n AC Adapter (3-prong) FZ-AA2202BM
STORAGE & MEMORY n 1GB RAM2,3
n LIND Car Adapter PA1220-3618
n 16GB flash2
n LIND Shutdown Timer CF-LINDLPT
– Up to 32GB additional with microSD card
n ToughMate A1 “X” Strap and Shoulder Strap TBCA1XSTP-P
DISPLAY n 10.1" XGA LED 1024 x 768

n ToughMate A1 Rotating Hand Strap
n Capacitive multi touch + digitizer daylight-readable screen with Shoulder Strap TBCA1HDSTP-P
– 500 nit
n ToughMate A1 Always-on Case TBCA1AONL-P
n ToughMate A1 Professional Portfolio TBCA1PFLIO-BLK-P
– Direct bonding technology
n ToughMate A1 Neoprene Sleeve TBCA1SLVNP-P
– Anti-reflective and anti-glare screen treatments
n Havis Docking Station:
– Ambient light, accelerometer and digital compass sensors
– Cradle with (2) USB, (1) Ethernet, (1) Power DS-PAN-501-P
– Automatic screen rotation
– Cradle with (2) USB, (1) Ethernet, (1) Power,
Audio n Integrated microphone (1) AC Adapter DS-PAN-502-P
n Integrated speaker – Cradle Only—No Charging or Communication DS-PAN-503
n On-screen volume and mute controls n Havis Desktop Stand DS-DA-218-P
n Barcode Reader Attachment FZ-A1-BCR
keyboard & input n Multi touch + digitizer screen n Barcode and Magstripe Readers Attachment FZ-A1-BCR-MSR
– Supports bare hand touch and gestures and electronic stylus pen n Replacement Stylus FZ-VNP001U
n 5 tablet buttons (Power, Back, Home, Menu, Dashboard) n Tether FZ-VNT001U
n Integrated stylus holder n 10.1" LCD Protector Film FZ-VPF001U
n On-screen QWERTY keyboard
Cameras n 2MP webcam with mic

n 5MP rear camera with auto focus and LED light
EXPANSION SLotS n MicroSD
interface n Docking connector 9-pin

n Micro-USB 2.0 4-pin
n Micro HDMI Type C
n Headphones/speaker Mini-jack stereo
wireless n Optional integrated 4G LTE mobile broadband with satellite GPS

n GPS (SiRFstarIII™)
n Wi-Fi 802.11a/b/g/n
n Wi-Fi hotspot router (up to 5 devices)4
Please consult your reseller or Panasonic representative before purchasing.
n Bluetooth® v2.1 + EDR (Class 2) Caution: Do not expose bare skin to this product when handling this unit in extreme hot or cold environments.
1
Panasonic has taken all steps available to secure the tablet for FIPS 140-2 (Level 2) Compliance. Panasonic is not
Power supply n Li-Ion battery pack (7.4V, typical 4770mAh, minimum 4590mAh) 2
responsible for any such security breaches caused by malicious attacks or improper feature implementation.
1GB = 1,000,000,000 bytes.
n Battery operation: 10 hours5 3
Total usable memory will be less depending upon actual system configuration.
n Battery 90% charging time: 3 hours on, 2.5 hours in sleep mode5
4
Only available when equipped with mobile broadband.
5
Battery performance features such as charge time and life span vary based on the conditions under which the
n AC Adapter: AC 100V-240V 50/60Hz computer and battery are used. Battery operation and recharge times will vary based on many factors, including
brightness, applications, features, power management, battery conditioning and other customer preferences. Battery
life with LCD brightness: 60 nit, Wi-Fi on while running a YouTube video.
power management n Sleep Mode 6
Length measurements do not include protrusions. Weight varies with options and digitizer pen.
7
Accessories and Integrated Options may vary depending on your configuration. Visit Panasonic website for more
security features n Embedded security processor for hardware encryption, trusted boot, root protection, accessories and details.
FIPS 140-2 (Level 2), AuthenTec™ QuickSec VPN, anti-virus protection (trial version),
support for RSA token authentication with 3rd party Cisco application, network proxy 1.800.662.3537
settings, password security (PIN, pattern, password), Kensington cable lock slot panasonic.com/toughpad/A1
SOLUTIONS FOR BUSINESS

Panasonic is constantly enhancing product specifications and accessories. Specifications subject to change without notice. Trademarks are property of their respective owners. ©2013 Panasonic Corporation of North
America. All rights reserved. Android, Google Play, YouTube and other marks are trademarks of Google Inc. The Android robot is reproduced or modified from work created and shared by Google and used according to
terms described in the Creative Commons 3.0 Attribution License. Toughpad FZ-A1 Spec Sheet_10/13
®
PowerFoil 115
Electrical characteristics
Measured at Standard test conditions Measured at Nominal Operating Cell Temperature
(STC; 25 °C cell temperature, insolation 1000 W/m2, (NOCT; ambient temperature 20 °C, insolation 800 W/m2,
AM 1.5) AM 1.5, 1 m/s wind speed)
Rated Power Pmax 116 W NOCT 45°C

Production tolerance of Pmax ±5 % Maximum Power Pmax 93 W
Rated Voltage Vmpp 17,2 V Voltage at max. power Vmpp 17,2 V
Rated current Impp 6,7 A Current at max. power Impp 5,4 A
Open circuit voltage Voc 24,8 V Open circuit voltage Voc 23,3 V
Short circuit current Isc 8,1 A Short circuit current Isc 6,6 A
Note: During the first weeks of operation, electrical output may exceed specified ratings. Power output may be higher by 15%,
operating voltage may be higher by 5 %, operating current may be higher by 10 %
Typical characteristics at varying irradiance levels Temperature coefficients (Tc)

(25 °C cell temperature, AM 1.5)
Tc of Pmax (% / °C) -0,20 %
9,0 Tc of Voc (% / °C) -0,31 %
8,0 Tc of Isc (% / °C) 0,02 %
7,0
6,0
Current (A)
5,0
4,0 Installation data
3,0
1.000 W/m2
2,0 Application class Class A at IEC 61730
500 W/m2
1,0
200 W/m2
Operating temperature - 40°C to + 85 °C
0,0
0 5 10 15 20 25 Maximum system voltage 500 V
Voltage (V) Maximum series fuse rating 13 A
General characteristics
Dimensions 5930x325x0.4 mm, depth at junction box 12 mm
Weight 1.3 kg
Cell type 28 amorphous silicon solar cells (5910X10 mm), connected in series
Front sheet fluorine polymer
Junction EPIC Solar Map
Connector quick-connect terminal (overmoulded)
Cable type Solar cable (4.0 mm2), length 325 mm
24,5 42
14
325
5930
115 - 001 EN
Flexible lightweight solar modules

®
PowerFoil 165
Electrical characteristics
Measured at Standard test conditions Measured at Nominal Operating Cell Temperature
(STC; 25 °C cell temperature, insolation 1000 W/m2, (NOCT; ambient temperature 20 °C, insolation 800 W/m2,
AM 1.5) AM 1.5, 1 m/s wind speed)
Rated Power Pmax 165 W NOCT 45°C

Production tolerance of Pmax ±5 % Maximum Power Pmax 132 W
Rated Voltage Vmpp 28 V Voltage at max. power Vmpp 28,0 V
Rated current Impp 5,9 A Current at max. power Impp 4,7 A
Open circuit voltage Voc 38,1 V Open circuit voltage Voc 36 V
Short circuit current Isc 6,5 A Short circuit current Isc 5,3 A
Note: During the first weeks of operation, electrical output may exceed specified ratings. Power output may be higher by 15%,
operating voltage may be higher by 5 %, operating current may be higher by 10 %
Typical characteristics at varying irradiance levels Temperature coefficients (Tc)

(25 °C cell temperature, AM 1.5)
Tc of Pmax (% / °C) -0,25 %
7,0 Tc of Voc (% / °C) -0,30 %
6,0 Tc of Isc (% / °C) 0,10 %
5,0
Current (A)
4,0
3,0 Installation data

2,0 1.000 W/m2
Application class Class A at IEC 61730
500 W/m2
1,0
200 W/m2
Operating temperature - 40°C to + 85 °C
0,0
0 5 10 15 20 25 30 35 40
Maximum system voltage 500 V
Voltage (V) Maximum series fuse rating 13 A
General characteristics
Dimensions 5930x325x0.4 mm, depth at junction box 12 mm
Weight 1.3 kg
Cell type 28 amorphous/microcrystalline tandem silicon solar cells (5910X10 mm), connected in series
Front sheet fluorine polymer
Junction EPIC Solar Map
Connector quick-connect terminal (overmoulded)
Cable type Solar cable (4.0 mm2), length 325 mm
24,5 42
14
325
5930
165 - 001 EN
Flexible lightweight solar modules

PowerFLEX™ BIPV 82/90/100W
Electrical Specifications
Capacity rating P max 100 W 90 W 82 W
Tolerance of Pmax % +10 / -7% +10 / -7% +10 / -7%
Module aperture area efficiency % 12.7% 11.4% 10.4%
Rated voltage V mpp 17.8 V 16.5 V 15.5 V
Rated current Impp 5.6 A 5.4 A 5.3 A
Open circuit voltage V oc 23.3 V 22.0 V 20.9 V
Short circuit current Isc 6.4 A 6.3 A 6.2 A
Note 1: Standard Test Conditions (STC): Cell Temperature at 25°C; Solar irradiance intensity of 1000 W/m ; AM1.5 solar reference spectrum (ASTM E892)
2
Note 2: Average efficiency is calculated using the 0.79 m2 aperture area of the module Note 3: Electrical parameters are +/-10% unless stated otherwise
Temperature Coefficients Low-Light Performance

Maximum power P max -0.43 % / °C Intensity Relative Efficiency
Maximum voltage V max -0.38 % / °C 1000 W/m2 100%
Open circuit voltage V oc -0.33 % / °C 500 W/m2 99%
Short circuit current Isc -0.03 % / °C 200 W/m 2
91%
Note: Relative to Standard Test Conditions (STC): Solar irradiance intensity of 1000 W/m2; Note: Relative to Standard Test Conditions
AM1.5 solar reference spectrum (ASTM E892) (STC): Cell Temperature at 25°C; AM1.5 solar
reference spectrum (ASTM E892)
Mechanical Specifications
Dimensions 2017 x 494 x <3 mm (79.3 x 19.4 x <0.12 in)
Weight 3.0 kg without adhesive (3.0 kg/m2 )
3.5 kg with adhesive (3.5 kg/m2 )
Junction Box - Top Mounted TE Connectivity SOLARLOKTM Micro Junction Box
Cables 4 mm2 dual rated with SOLARLOKTM connectors
Junction Box - Bottom Mounted MC-JM/FM (MC4 Connectors)
Front Sheet Non-stick E T F E
Solar Cells 36 CIGS cells (210 x 100 mm)
Adhesive ADCO HelioBondTM PVA 600BT butyl mastic
Hot Spot Protection Bypass diodes at each cell; 1 at junction box
Materials Lead free and exempt from RoHS requirements
Maximum Series Fuse Rating 10 Amp
Color Options Black (B) or White (W) backsheet behind cells
Operating Conditions
Temperature Range -40°C to + 85°C
Maximum System Voltage 1000VDC IEC, 600VDC UL
Certifications and Warranty*

EN 61646, EN 61730, UL 1703 Also available in
Materials and workmanship - 5 years 4.0m x 0.5m or 6.0m x 0.5m modules
Power output - 25 years (90% @ 10 yrs; 80% @ 25 yrs) Limited Warranty
*Contact GSE for complete warranty terms
Global Solar ® is a leading manufacturer of Copper Indium Gallium diSelenide (CIGS) thin-film solar on a flexible substrate. With a plant in Tucson
(Arizona, USA), Global Solar ® operates with a total of 40MW of production capacity. An average cell efficiency above 12.5% makes the company the
world leader in CIGS efficiency on flexible substrate in large scale production. Sold worldwide in multiple applications, including flexible laminates, solar
shingles, glass modules and portable chargers.
Call Global Solar ® to find out if PowerFLEX™ BIPV is right for you.
Global Solar ® Energy, Inc.

8500 South Rita Road
Tucson, Arizona, 85747, USA
Phone: +1 (520) 546-6313
Fax: +1 (520) 546-6318
A COMPANY OF
[email protected]
www.globalsolar.com
© Global Solar ® Energy, Inc. Global Solar ®, PowerFLEXTM, and PowerFLEX Technology ® are trademarks of Global Solar ® Energy, Inc.
Doc No. 1000397 rev D
SOLAR CHARGE CONTROLLERS
Steca Solarix MPPT

MPPT 1010, MPPT 2010
Steca Solarix MPPT is a solar charge controller with Maximum Power
Point Tracking. It is specially designed to work with all established
module technologies and is optimized for solar systems with module
voltages higher than the battery voltage. The Steca Solarix MPPT
is especially qualified in combination with grid tied solar modules.
The advanced MPP-tracking algorithm from Steca assures that
the maximum available power of the solar generator is charged
to the batteries. The Steca Solarix MPPT with latest technology
ensures full performance in all conditions, a professional battery
care combined with modern design and excellent protection.
10 A...20 A
Product features
∙ Maximum Power Point Tracker (MPP tracker)
∙ Voltage and current regulation
∙ PWM control
∙ Current compensated load disconnection
68
∙ Automatic load reconnection
∙ Temperature compensation
∙ Monthly maintenance charge 187
ø5
4x
Electronic protection functions
∙ Overcharge protection
∙ Deep discharge protection
∙ Reverse polarity protection of module, load and battery
∙ Reverse polarity protection by internal fuse
153
115
∙ Automatic electronic fuse
∙ Short circuit protection
∙ Overvoltage protection at module input
∙ Open circuit protection without battery
15
∙ Reverse current protection at night
177 5
∙ Overtemperature and overload protection
∙ Battery overvoltage shutdown
MPPT 1010 MPPT 2010
Characterisation of the operating performance
Displays
∙ Multifunction LED display System voltage 12 V (24 V)
∙ Multi-coloured LED Nominal power 125 W (250 W) 250 W (500 W)
∙ 5 LEDs show operating states Max. efficiency > 98 %
~ for operation, state of charge, fault messages Own consumption 10 mA
DC input side
Options 15 V (30 V) < Vmodule 15 V (30 V) < Vmodule
∙ Night light function pre-set in the factory or adjustable MPP voltage
< 75 V << 100 V
via Steca PA RC 100 Open circuit voltage solar module 17 V ... 75 V 17 V ... 100 V
∙ Parameterisation of function values via Steca PA RC 100 (at minimum operating temperature) (34 V … 75 V) (34 V … 100 V)**
∙ External temperature sensor Module current 9A 18 A
DC output side
Certificates Charge current 10 A 20 A
∙ Compliant with European Standards (CE) Load current 10 A
∙ RoHS compliant
End of charge voltage* 13.9 V (27.8 V)
∙ Made in Germany
Boost charge voltage* 14.4 V (28.8 V)
∙ Developed in Germany
∙ Manufactured according to ISO 9001 and ISO 14001
Equalisation charge* 14.7 V (29.4 V)
Reconnection voltage (LVR)* 12.5 V (25 V)
Deep discharge protection (LVD)* 11.5 V (23 V)
Operating conditions
η [%]
100 Ambient temperature -25 °C … +40 °C
99 Fitting and construction
98 Set battery type liquid (adjustable via Steca PA RC100)
24 V
97 Terminal (fine / single wire) 16 mm2 / 25 mm2 - AWG 6 / 4
96 12 V Degree of protection IP 32
95 Dimensions (X x Y x Z) 187 x 153 x 68 mm
94 Weight approx. 900 g
93
* see options Technical data at 25 °C / 77 °F
92 **CAUTION: If an open circuit voltage of more than 100 V is supplied to the connected
91 solar module, the controller will be destroyed. When selecting the solar module, it is
important to bear in mind that the open circuit voltage should never exceed 100 V over
90 the entire working temperature range. When using solar modules with a maximum open
25 50 75 100
circuit voltage of between 75 and 100 V (over the entire temperature range), all installa-
Power [% of the rated power] tion steps must be carried in accordance with protection class II.
Areas of application: Steca PA TS10 Steca PA RC100

External Remote control
480 W temperature sensor
99
The charge controller in Matlab
1 function [ P_battery , P_dump , P_deficit ]= chargecontroller ( P_load , P_module ,

SOC_battery )
2 % P_battery = power into the battery :
3 % If P_battery is negative , power is flowing from the battery
4
5 % P_dump = power dumped because of full battery ;
6 % Dump is not a physical system component ,
7 % but just modeled to analyze the system performance .
8
9 % P_deficit = Power not fed to load because of empty battery ;
10 % again , just modeled to analyze the system performance .
11
12 % Charge controller working principle :
13 if SOC_battery <20 && P_module<P_load
14 P_battery=P_module ;
15 P_dump =0;
16 P_deficit=P_load ;
17 elseif SOC_battery >80 && P_module>P_load
18 P_battery =0;
19 P_dump=P_module−P_load ;
20 P_deficit =0;
21 else
22 P_battery=P_module−P_load ;
23 P_dump =0;
24 P_deficit =0;
25 end

100 Appendices

Bibliography
[1] International Energy Agency, “Technology Roadmap: Solar Photovoltaic Energy,” tech.
rep., International Energy Agency, May 2010.
[2] V. Tyagi, N. a.a. Rahim, N. Rahim, and J. a./L. Selvaraj, “Progress in solar PV tech-
nology: Research and achievement,” Renewable and Sustainable Energy Reviews, vol. 20,
pp. 443–461, Apr. 2013.
[3] “Eco-Economy Indicators - Solar Power | EPI.” [url: http://www.earth-

policy.org/indicators/C47, visited on: 25/09/2014].
[4] M. Humbert, “Technology and Workforce : Comparison between the Information Revo-
lution and the Industrial Revolution,” no. December, 2007.
[5] “Dutch smartphone penetration hits 72% in Q3.” [url:

http://www.telecompaper.com/news/dutch-smartphone-penetration-hits-72-in-q3–
973995, visited on: 24/07/2014].
[6] “Information kiosk.” [url: http://www.turbophoto.com/Free-Stock-

Images/Images/Information Kiosk.jpg, visited on: 25/07/2014].
[7] “TU Delft: Photovoltaic Materials and Devices.” [url: http://www.ewi.tudelft.nl/en/the-

faculty/departments/electrical-sustainable-energy/photovoltaic-materials-and-devices/,
visited on: 26/07/2014].
[8] “EPSI Conference 2012 - Conference Venue.” [url: http://epsi.eu/conference_2012/epsi-

conference-2012-conference-venue/, visited on: 15/08/2014].
[9] L. Entman, “Vu solar charging station,” 2012. [url:

http://news.vanderbilt.edu/2012/05/solar-charging-stations/, visited on: 15/08/2014].
[10] “Mit soft rocker.” [url: http://www.designboom.com/design/mit-soft-rocker-solar-

powered-sun-lounger/, visited on: 25/07/2014].

102 Bibliography
[11] “Rugged Display | Rugged Monitor | Stainless Vehicle Touch LCD-AIS.” [url:
http://www.aispro.com/products/rugged-touch-monitor-displays, visited on:
15/09/2014].
[12] “Rugged PC Review.com - Rugged Tablet PCs: Handheld Algiz 10x.” [url:
http://www.ruggedpcreview.com/3_slates_handheldus_algiz10x.html, visited on:
15/09/2014].
[13] “Toughpad - Computer Product Solutions | Panasonic Business.” [url:

http://business.panasonic.nl/computeroplossingen/producten-en-diensten/panasonic-
toughbook/toughpad, visited on: 15/09/2014].
[14] “Panasonic ToughPad FZ-G1 - Prijzen en Specificaties - Tweakers.” [url:

http://tweakers.net/product/347508/panasonic-toughpad-fz-g1/, visited on:
15/09/2014].
[15] “Panasonic ToughPad FZ-A1 WiFi Zilver - Prijzen - Tweakers.” [url:

http://tweakers.net/pricewatch/335018/panasonic-toughpad-fz-a1-wifi-zilver.html,
visited on: 15/09/2014].
[16] “Panasonic Fully Rugged Tablet | Business Tablet | Toughpad FZ-A1.” [url:
http://www.panasonic.com/business/toughpad/us/best-android-rugged-tablet-
overview.asp, visited on: 15/09/2014].
[17] “Dined.” [url: http://dined.io.tudelft.nl/dined/full, visited on: 16/09/2014].
[18] “Google Maps.” [url: http://maps.google.com, visited on: 17/09/2014].
[19] K. . Zonen, “Solar Radiation Measurements for Solar Energy Applications.” [url:
http://www.kippzonen.com/?download/415212/Solar+Energy+Guide.aspx, visited on:
26/09/2014].
[20] N. Narayan, “Solar Charging Station for Light Electric Vehicles,” 2013.
[21] I. Reda and A. Andreas, “Solar Position Algorithm for Solar Radiation Applications
Solar Position Algorithm for Solar Radiation Applications,” no. January, 2008.
[22] “Making Use of TMY Data.” [url: http://www.pveducation.org/pvcdrom/properties-of-

sunlight/making-use-of-TMY, visited on: 19/09/2014].
[23] “Simple Sandia Sky Diffuse Model - PV Performance Modeling Collaborative.”

[url: http://pvpmc.org/modeling-steps/incident-irradiance/plane-of-array-poa-
irradiance/calculating-poa-irradiance/poa-sky-diffuse/simple-sandia-sky-diffuse-model/,
visited on: 19/09/2014].
[24] D. T. Reindl, W. A. Beckman, and J. A. Duffle, “Diffuse fraction correlations,” Solar

Energy, vol. 20, no. I, 1990.
[25] D. Reindl, W. Beckman, and J. Duffie, “Evaluation of hourly tilted surface radiation
models,” Solar Energy, vol. 45, pp. 9–17, Jan. 1990.

103
[26] P. Loutzenhiser, H. Manz, C. Felsmann, P. Strachan, T. Frank, and G. Maxwell, “Em-

pirical validation of models to compute solar irradiance on inclined surfaces for building
energy simulation,” Solar Energy, vol. 81, pp. 254–267, Feb. 2007.
[27] P. Ineichen, O. Guisan, and R. Perez, “Ground-reflected radiation and albedo,” Solar
Energy, vol. 44, pp. 207–214, 1990.
[28] A. McEvoy, T. Markvart, and L. Castaner, Practical Handbook Of Fundamentals and
Applications. Academic Press, 2nd ed., 2012.
[29] “CTB3300WCx - GWC 2 - Evaporation.” [url:
https://www.youtube.com/watch?v=AFdhlVEQTTA, visited on: 09/09/2014].
[30] “Effect of Temperature | PVEducation.” [url: http://www.pveducation.org/pvcdrom/solar-
cell-operation/effect-of-temperature, visited on: 03/04/2014].
[31] M. K. Fuentes, “A Simplified Thermal Model for Flat-Plate Photovoltaic Arrays,” 1987.
[32] P. Trinuruk, C. Sorapipatana, and D. Chenvidhya, “Estimating operating cell temper-
ature of BIPV modules in Thailand,” Renewable Energy, vol. 34, pp. 2515–2523, Nov.
2009.
[33] J. Duffie and W. Beckman, Solar Engineering of Thermal Processes. John Wiley & Sons,
Inc., 2006.
[34] M. Fuentes, G. Nofuentes, J. Aguilera, D. Talavera, and M. Castro, “Application and
validation of algebraic methods to predict the behaviour of crystalline silicon PV modules
in Mediterranean climates,” Solar Energy, vol. 81, pp. 1396–1408, Nov. 2007.
[35] “Hot Spot Heating | PVEducation.” [url: http://pveducation.org/pvcdrom/modules/hot-
spot-heating, visited on: 28/09/2014].
[36] “Implement generic battery model - Simulink.” [url:
http://www.mathworks.nl/help/physmod/sps/powersys/ref/battery.html, visited
on: 01/10/2014].
[37] Planning and Installing Photovoltaic Systems: A Guide for Installers, Architects and
Engineers. London: Earthscan, second ed., 2008.
[38] Hoppecke, “Installation , commissioning and operating instruc-
tions for vented stationary lead-acid batteries,” 2013. [url:
http://www.hoppecke.com/content/download/brochures/rp/TechnicalDocumentation/...
.../Montagehandbuch_verschl_en.pdf, visited on: 01/10/2014].
[39] Gaston, “Vrla battery general catalogue.” [url:
http://www.gaston.com.hk/pdf/VRLA_BATTERY_GENERAL_CATALOGUE.pdf,
visited on: 17/10/2014].
[40] Solaris-Store, “Régulateur solaire steca solarix mppt 1010.” [url: http://www.solaris-
store.com/39-regulateur-solaire-steca-solarix-mppt-1010.html, visited on: 24/10/2014].
[41] G. S. Hanergy, “Powerflex™bipv - 82/90/100w.” [url:
http://www.globalsolar.com/downloads/PowerFLEX_2.pdf, visited on: 29/08/2014].

104 Bibliography
[42] Fasttech, “Star go st-07 2.1a 2*usb car cigarette lighter charger adapter.”
[url: http://www.fasttech.com/product/1558007-star-go-st-07-2-1a-2-usb-car-cigarette-
lighter, visited on: 24/10/2014].
[43] “RS 12V Lead Acid Rechargeable Battery, AGM Construction, 150Ah.” [url: http://nl.rs-
online.com/web/p/lead-acid-rechargeable-batteries/7270442/, visited on: 24/10/2014].

Glossary
List of Acronyms
PVMD Photovoltaic Materials and Devices
TU Delft Delft University of Technology
PV photovoltaic
STC Standard Test Conditions
NOCT normal operating cell temperature
INOCT installed normal operating cell temperature
GHGs greenhouse gases
IEA International Energy Agency
ICT Information and Communications Technology
EEMCS Electrical Engineering, Mathematics and Computer Science
MIT Massachusetts Institute of Technology
KNMI Royal Netherlands Meteorological Institute
SPA Sun Path Algorithm
CEG Civil Engineering and Geosciences
FMRE Facility Management & Real Estate
NOCT Nominal Operating Cell Temperature
INOCT Installed Nominal Operating Cell Temperature
SRE standard reference environment
a-Si amorphous silicon

106 Glossary
CIGS Copper Indium Gallium diSelenide
µc-Si microcrystalline silicon
SOC state of charge
AC Alternating Current
DC Direct Current
MPPT Maximum Power Point Tracking
LVD Low Voltage Disconnect
LVR Low Voltage Reconnect
LLP loss-of-load-probability
PUE percentage of unused energy
DOD depth-of-discharge
IDE Industrial Design Engineering
VRLA Valve Regulated Lead-Acid
DEMO Electronic and Mechanical Support Division

107
List of Symbols
CO2 carbon-dioxide
Voc open circuit voltage
Isc short circuit current
Pmpp maximum power point
Vmpp maximum power point voltage
GHI Global Horizontal Irradiance
DN I Direct Normal Irradiance
DHI Diffuse Horizontal Irradiance
Asun azimuth angle of the sun
asun altitude angle of the sun
A azimuth angle
Amodule azimuth angle of the module
θscreen screen tilt angle
θplane tilt angle of the plane
Aplane azimuth angle of the plane
AOI angle of incidence
SV F sky view factor
H height of the highest point of the curvature
F footprint
L length of the curvature
V voltage
J current density
κ temperature coefficient of Pmpp

108 Glossary

V Weeda MscThesis

Uploaded by

Copyright:

Available Formats

V Weeda MscThesis

Uploaded by

Document Information

Original Title

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

V Weeda MscThesis

Uploaded by

Copyright:

Available Formats

Solar Powered Infotainment Spot

The design, feasibility study and fabrication of an au-

Photovoltaic Materials and Devices

Master of Science Thesis

For the degree of Master of Science in Sustainable Energy Technology

October 31, 2014

Faculty of Applied Sciences · Delft University of Technology

Dated: October 31, 2014

Prof. Dr. M. Zeman

Master of Science Thesis V. Weeda

V. Weeda Master of Science Thesis

2 Design & Location 7

Master of Science Thesis V. Weeda

5 Building the system 65

6 Conclusions and Recommendations 87

V. Weeda Master of Science Thesis

1-1 Annual PV electricity production. . . . . . . . . . . . . . . . . . . . . . . . . . . 2

3-1 The components of irradiance on a tilted surface. . . . . . . . . . . . . . . . . . 20

Master of Science Thesis V. Weeda

3-11 Diffuse irradiation on a tilted plane, location-specific SV F . . . . . . . . . . . . 31

4-1 Definition of variables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

5-1 Topology of the system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

V. Weeda Master of Science Thesis

5-12 Definition of the energy flows in/from the charge controller. . . . . . . . . . . . 76

Master of Science Thesis V. Weeda

V. Weeda Master of Science Thesis

2-1 Physical dimensions of Dutch adults. . . . . . . . . . . . . . . . . . . . . . . . . 12

3-1 Irradiation on a horizontal surface in 2012 and 2013. . . . . . . . . . . . . . . . 23

4-1 DC yield prediction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

Master of Science Thesis V. Weeda

V. Weeda Master of Science Thesis

1-1 Energy Crisis

1-2 Information Age

Master of Science Thesis V. Weeda

Figure 1-1: Annual PV electricity production, adapted from: [3].

1-3 Motivation of the thesis

V. Weeda Master of Science Thesis

Figure 1-2: Information kiosk on a train station in Perth, Australia [6].

Figure 1-3: The location: TU Delft campus [8].

Master of Science Thesis V. Weeda

1. Demonstrating what is currently possible in the PV technology with the implementation

• Practical information on TU Delft and the faculty Electrical Engineering, Mathematics

1-3-2 Scope of the thesis

V. Weeda Master of Science Thesis

1-4 Research questions

Master of Science Thesis V. Weeda

1-5 Thesis outline

Chapter 2 - Design & Location

Chapter 3 - Energy Resources

Chapter 5 - System modeling

Chapter 6 - Conclusions and Recommendations

V. Weeda Master of Science Thesis

Design & Location

Master of Science Thesis V. Weeda

V. Weeda Master of Science Thesis

(a) Concept A: The umbrella (b) Concept B: The lectern

Master of Science Thesis V. Weeda

V. Weeda Master of Science Thesis

Figure 2-2: The Panasonic Toughpad FZ-A1 [16].