Design and Installation of A Grid-Connected PV System
Design and Installation of A Grid-Connected PV System
Design and Installation of A Grid-Connected PV System
PV System
Supervisor
Georgi Hristov Yordanov
Abstract
As the worlds electrical demand is being more covered by photovoltaic (PV) systems, Germany
installed 3.3 GW of the worlds PV systems in 2013 compared to 0.6 MW being installed in
Norway. Irradiation measurements done for southern Norway indicates that the conditions for
more PV systems being installed are good when comparing irradiation in Grimstad to Hamburg.
This thesis addresses the initial steps of planning and building a PV system to analysing how
the inverter handles the varying irradiance in southern Norway.
To investigate the opportunities in Grimstad, a Theia HE-t 3.8 string inverter from Eltek was
installed along with 23 de-rated PV modules to constitute a PV system of 2.116 kWp . The PV
system was estimated to have a system loss of 20.4 % and a yearly production of 2088 kWh with
an average performance ratio of 80 %, which is considered to be good for a small PV system
with aged PV modules. From when the system started commissioning to when this thesis was
finalized, approximately one month of measurements was recorded, providing information on
how the inverter operated on clear and cloudy days as well as days with overirradiance.
Results indicate that the inverter efficiency drops during early and late operating hours on
sunny days, as well as during midday on cloudy days. This was found to be a result of the
inverter using two dierent algorithms for low and high irradiance. When the inverter switches
algorithms, it could operate at a not optimal efficiency for 10 minutes up to two hours dependent
of the available irradiance, the inverter will then operate at a high efficiency for as long as it
is operating with one specific algorithm. One day with overirradiance reviled that the inverter
responds rapidly to sudden irradiance changes and were able to produce 2.8 kWAC at irradiance
close to 1.5 suns. It therefore has good preconditions for handling varying irradiance while
operating at an average high efficiency. For the one month of operation the inverter managed
to produce 313 kWh, which is more than was estimated for May. With a lack of more months
to compare with and from the results recorded for May, the PV system has good prospects for
future operation and to produce the expected 2088 kWh/year or more.
Preface
This is a masters thesis submitted in fulfilment of the requirements for the degree of Master
in Technology in Renewable Energy at the department of engineering sciences, University of
Agder. We would like to thank our supervisor Dr Georgi Hristov Yordanov for good comments,
explanations, discussions and help at all aspects during the progress of this thesis.
We would like to thank Bjrn Lindemann, Stein Bergsmark and Kristen Leifsen from UiA
who were essential for obtaining approvals, providing electrical services and components, and
securing a good communication with Rolf H
akan Josefsen from Agder Energi regarding the grid
connection of the inverter. We would also like to thank Kjetil Boysen and Espen Kristensen
from Eltek for advices regarding the inverter, providing the PC tool (Theia Analyzer) and the
python script for data logging.
An extra thanks is sent to Stein Bergsmark for giving valuable constructive criticism on the
content and language of the thesis as well as project funding, giving us the possibility to buy
the inverter, execute the grid connection and write a good thesis.
Last but not least we would like to thank Prof. Tor Oskar Stre from UiA who provided part of
the funding for the projects implementation and Steve Schading from UiA who advised on the
selection of proper fasteners for mounting the PV modules in order to avoid galvanic corrosion.
Grimstad, June 2014.
John Christer Sivertsen
Petter Syland
ii
Contents
Abstract
Preface
ii
Contents
iii
List of Figures
vi
List of Tables
ix
Abbreviations
1 Introduction
1.1 Motivation . . . . . . . . . .
1.2 Project Description . . . . . .
1.2.1 Goals and Objectives .
1.2.2 Key Assumptions . . .
1.3 Thesis Outline . . . . . . . .
2 PV
2.1
2.2
2.3
.
.
.
.
.
.
.
.
.
.
Inverters
The Basic Inverter . . . . . . . .
Inverter Transformers . . . . . .
Grid-connected Inverters . . . . .
2.3.1 MPPT . . . . . . . . . . .
2.3.2 Power Rating of Inverters
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
1
2
2
4
4
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
6
6
9
10
11
11
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
14
14
15
17
18
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
20
20
22
23
26
3 Grid Connection
3.1 The Plus Customer . . . . .
3.2 Grounding Systems . . . . .
3.3 Islanding . . . . . . . . . .
3.4 Grid Application Approach
.
.
.
.
.
.
.
.
.
.
.
.
iii
4.3
4.4
4.5
4.6
4.7
4.8
4.9
Ambient Temperature . . . . .
Voltage Drop Along the Cables
Wind load . . . . . . . . . . . .
Temporary Shading . . . . . .
Corrosion . . . . . . . . . . . .
Grounding . . . . . . . . . . . .
Fire Hazard . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
27
29
30
31
33
34
36
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
37
37
40
41
41
42
43
45
46
46
47
48
48
49
49
55
61
64
64
65
66
66
67
68
68
68
68
69
70
71
71
72
76
77
77
79
80
81
83
iv
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5.8
5.7.6 Commissioning . . . . . . .
5.7.7 Thermal Imaging . . . . . .
Inverter Performance . . . . . . . .
5.8.1 Analysis of Operating Data
5.8.2 Performance Ratio . . . . .
5.8.3 Energy Harvest . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 83
. 86
. 88
. 88
. 103
. 104
6 Conclusion
107
Bibliography
110
119
122
C Specifications of Inverters
127
129
E PV Module Ranking
135
137
139
141
143
J Technical Guidelines (Attachment 3 from the Agreement between Agder Energi Nett and UiA)
147
K Documentation (Attachment 5 from the Agreement between Agder Energi
Nett and UiA)
154
L General Agreement (Main Document between Agder Energi Nett and UiA)159
M Datasheet of the PV modules
163
165
167
List of Figures
1.1
2.1
2.2
2.3
2.4
2.5
6
7
8
8
2.6
2.7
3.1
3.2
3.3
3.4
4.1
4.2
4.3
4.4
9
11
13
15
16
16
18
21
24
27
5.1
43
. . . . . . . . . . . . . . . . . . . . . . .
28
28
30
31
32
32
35
36
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
5.10
5.11
5.12
5.13
5.14
5.15
5.16
5.17
5.18
5.19
5.20
5.21
5.22
5.23
5.24
5.25
5.26
5.27
5.28
5.29
5.30
5.31
5.32
5.33
5.34
5.35
(a) A voltage dip caused by a short circuit in the grid (b) A voltage swell caused
by a disconnection of a large load. . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Illustrates how some of the arrays power cannot be processed because it is outside
the MPPT range of the inverter. . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
PV array configuration as it was installed on the roof, showing the order and
polarity of the PV modules as well as the four elongated string cables. . . . . . . 56
Correct and incorrect wiring of the string cables with regards to minimizing the
wiring loop. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2-line diagram of the PV system. . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
The relative efficiency is calculated with respect to STC, covering the most measured irradiances over Grimstad in 2011. . . . . . . . . . . . . . . . . . . . . . . . 65
Soil and dirt de-rating examples of the PV system. . . . . . . . . . . . . . . . . . 66
Illustration of some of the shading factors. . . . . . . . . . . . . . . . . . . . . . . 67
The temperature for module 30 changes throughout the testing time and is illustrated over the amount of measurements done for this specific module. . . . . . . 71
Set-up used when testing the available PV modules with ESL-Solar 500DS. . . . 72
I-V curves from module 30. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Histogram of STC power values from module 30, median values are used when
ranking the PV modules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Faults found during visual inspection. . . . . . . . . . . . . . . . . . . . . . . . . 81
Configuration of the PV array after the mechanical installation. . . . . . . . . . . 82
The inverter installed on the wall along with the junction box and wire connections. 82
Temperature and irradiance measurements on 27 April 2014, the highest power
was recorded at 12:00. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Thermal images of the inverter taken when operating at a high power [19]. . . . . 86
Thermal images of the array on a sunny day indicating that bird droppings and
junction boxes are a heat source. . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Thermal inspection of the cable MC3 connectors. It is seen that no heat is evolved
at this point. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Irradiance in Grimstad and inverter power on 2 May 2014. . . . . . . . . . . . . . 89
Inverter voltage on 2 May 2014. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Module Temperature on 2 May 2014. 459, Titan and New mc-Si are
dierent measuring points on the array. . . . . . . . . . . . . . . . . . . . . . . . 91
Inverter current with the oscillation at low irradiance on 2 May 2014. . . . . . . 92
Irradiance, inverter current and efficiency on 6 May 2014. . . . . . . . . . . . . . 93
Inverter power on 6 May 2014. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
A drop in power, illustrated with the efficiency on 2 May 2014. . . . . . . . . . . 95
Input current and voltage are compared on two days with dierent sky conditions. 96
Irradiance on 18 May 2014, indicating varying irradiance at midday and in the
afternoon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Sky conditions during overirradiance on 9 May 2014. . . . . . . . . . . . . . . . . 97
Overirradiance over Grimstad of almost 1.5 suns on 9 May 2014, the time values
are approximate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Power and current produced during overirradiance on 9 May 2014. . . . . . . . . 98
Inverter and Current efficiencies on 9 May 2014. . . . . . . . . . . . . . . . . . . 99
Inverter efficiency of Theia HE-t compared to SMA 2500 on a sunny day in
Grimstad and Klamath Falls, respectively. . . . . . . . . . . . . . . . . . . . . . . 100
Array temperatures on 5 and 6 May. . . . . . . . . . . . . . . . . . . . . . . . . . 104
vii
5.36 A screenshot of Theia Analyzer showing the inverter power output during the last
7 days, from 19 to 25 May. It indicates a varying power output due to irradiance,
a very low production on 21 May and a production of 1828 W on 24 May. . . . 105
5.37 A screenshot of Theia Analyzer showing a) The total energy produced until 13:00
on 26 May of 8661 Wh, b) Energy produced every day in May, giving a total
energy of 238 kWh this month, c) Energy produced so far this year, 313 kWh
has been produced since the inverter started commissioning. . . . . . . . . . . . . 106
viii
List of Tables
4.1
4.2
4.3
4.4
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
5.10
5.11
5.12
5.13
5.14
5.15
5.16
5.17
Average sum of global irradiation per square metre received by the modules of the
h
given system ( kW
). Mounting position towards south, with optimum inclination.
m2
Radiation database: Classic PVGIS (based on ground measurements. . . . . . . .
Albedo values for selected surfaces. . . . . . . . . . . . . . . . . . . . . . . . . . .
Average monthly global irradiation per square metre received by the modules of
h
a given system ( kW
). Mounting position is 7 east from south, with 39 1 tilt
m2
angle from the horizontal. Location: Agder Photovoltaic Lab, University of Agder.
Galvanic series, showing metallic elements and its corresponding electrode potential.
22
23
26
33
ix
Abbreviations
AC
Alternating Current
BIPV
Building-Integrated Photovoltaics
CAN
CCC
DC
Direct Current
DSP
FoL
GSES
HF
High Frequency
IEA
IGBT
JRC
LF
Low Frequency
MOSFET
MPP
MPPT
NVE
PCC
PR
Performance Ratio
PVGCS
PVGIS
PWM
RCD
Residual-Current Device
RMS
STC
THD
UiA
University of Agder
PR
Performance Ratio
xi
Chapter 1
Introduction
In this chapter, the background and motivation of this thesis is explained. The project description is presented, with goals, objectives, key assumptions and an outline of the thesis.
1.1
Motivation
By year 2020 the European Union aims to get 20 % of its energy from renewable sources [1]. One
of the technologies helping EU reach this goal is photovoltaic (PV) power generation, which has
grown rapidly in Europe due to governmental support. PV systems are a competitive alternative
to other renewable sources currently because of subsidies and in some countries like Germany
they can even compete with building new fossil fuel plants. The International Energy Agency
(IEA) PV power systems programme registered that approximately 39.9 GW of PV systems
was installed over the world throughout 2013, which was an increase of 7.9 GW compared to the
past two years. The PV installations in Europe decreased that year, installing only 10.3 GW
compared to 17 GW in 2012. China was the country installing most PV systems in the world
in 2013 with its 11.3 GW compared to Germany with 3.3 GW and Norway with 0.6 MW [2].
PV systems in Norway have traditionally been mostly limited to residential housings and boats
where connection to the electrical grid is difficult. Grid connected systems have not been widely
developed in Norway due to both technical and economic barriers with no support from the
government. Getting concession to feed energy into the grid has been a relatively complicated
process suited for professional electricity producers. However, in 2010 NVE made an exception
to the regulations to make it easier for customers to feed energy into the grid. In the near
1
Chapter 1. Introduction
future, it is expected that Norway will have a complete framework of regulations making it
more beneficial and easier to buy and install a grid-connected PV system, which may give the
required boost necessary to create a more healthy and mature PV market in Norway.
1.2
Project Description
The purpose of this master project is to establish and learn the necessary skills and experience
required to build a grid-connected PV system. The project involves planning and designing;
decide array size, number of strings, circuit breakers and choose modules based on I-V curve
measurements; install and connect the PV system to the grid; perform commissioning; operate
the PV system and analyse operational data. PV modules from the former National Renewable
Energy Centre (Energiparken) will be reused in this project. Some of these PV modules are
already mounted on the planned installation site and will be complemented with additional
modules if the I-V-characteristics are of good quality.
1.2.1
Goals
The goals of this project are to:
Chapter 1. Introduction
Install and if possible (depending on I-V curve measurements) reuse modules already
present at the installation site. Wire modules in one or two strings and connect to inverter
and grid with help from the certified sta of UiA.
Perform commissioning of the PV system.
Analyse inverter performance and operational data.
Objectives
In order to accomplish the goals of this project the following objectives are needed.
Study the present and the expected Norwegian regulations for feeding power into the AC
grid and study the policy of the local grid owner Agder Energi.
Study the local market for PV inverters and determine the most suitable model(s). Detailed logging on the AC and DC sides is required and the inverter should have a power
rating of about 3.5 kW .
Compare the local tilted irradiation from the years 2011-2013 to the long-term prediction
of PVGIS.
Measure the current-voltage (I-V) characteristics of the available PV modules; identify the
modules with drastically reduced current; choose modules with the best characteristics for
the PV array.
Size the PV system with respect to the power rating of the inverter; maximum system
voltage of the modules; the no-exceed voltage of the inverter; and available space at the
chosen installation site.
Determine the maximum wind load expected on the PV modules.
Install PV modules, connect them in strings and label each of the modules according to
its position in the PV array.
Analyse operational data, such as: energy harvest, current and output power.
Investigate how the inverter performs throughout the day at dierent irradiance conditions.
Chapter 1. Introduction
If Time Allows
The following objectives will be considered if time allows.
If there are days with extreme overirradiance events, and if the data logger oers very fine
temporal resolution, assess the inverter performance during the strongest bursts.
1.2.2
Key Assumptions
In order to complete the goals and objectives this project is based on the following assumptions:
Permission will be granted from Agder Energi, the Facultys administration and the building owner J.B. Ugland Eiendom to connect the PV system to the grid.
System parts are provided and installed in time.
Access to necessary software and licences.
The link between the inverter and the AC grid will be made timely by certified sta of
UiA.
1.3
Thesis Outline
The objectives state that an inverter has to be chosen for the PV system. Chapter 2 will
explain the basics on how a full-bridge voltage source inverter works and the dierent kinds of
grid connected inverters existing on the market.
Chapter 3 explains the dierent principles of grid connection, like islanding, microgrid, phasecurrent, etc. Then, a review of the application process for connecting the PV system to a local
utility grid is conducted and explained stepwise.
Chapter 4 presents an introduction to solar irradiance and the basic principles of measuring irradiance. Estimated irradiation from PVGIS is studied and compared with local measurements in
Grimstad. External factors like temperature, wind load, temporary shading, corrosion, grounding and fire hazard are also explained in this chapter.
Chapter 1. Introduction
Chapter 5 is the most important part of the thesis and includes the results, discussion and
issues. The selected inverters are compared and discussed before a decision is made for what
inverter to select as well as the local regulations for grid connection of a private production plant
given by Agder Energi, which the inverter must comply with. The information is evaluated and
cited in Section 5.2 and 5.3, giving restrictions for the required inverter power quality and
disconnection times in case of a grid fault. Factors with a direct influence on the performance
and causing losses in the PV system are discussed in Section 5.5. Chapter 5 also contains a
full guideline on how to design a PV system from the array- and cable sizing to calculating
losses and system protection. In Section 5.6 it is explained how the available PV modules
from the former National Renewable Energy Centre (Energiparken) are connected and tested.
The I-V curves for each module are measured by using a computer-controlled electronic load.
Prior to the mechanical and electrical installation process, a list of the best PV modules is
made by analysing the measured I-V curves. The installation process is presented in Section
5.7 and covers the mechanical and electrical installation together with the inverter installation,
commissioning of the PV system and thermal imaging. In addition, operational data from the
PV system with one-second resolution from a few days of operation are analysed and discussed
in Section 5.8 explaining how the inverter is performing at low-, high- and over-irradiance.
How to design a PV system is explained in detail in the Australian manual Grid-connected
PV systems Design and Installation [3] and the German manual Planning and Installing
Photovoltaic Systems, a guide for installers, architects and engineers [4]. These manuals are
used throughout the whole project as guidelines to design the array size, cable size and protection
devices (lightning and earthing). The German manual was at some times hard to come by from
the library, since it was often on loan to other groups. The Australian manual is therefore
mostly used. This did however, not cause an issue since the two manuals are very alike and the
Australian manual is even referring to the German manual. The largest dierence between the
two is the extent they explain methods and dierent concepts. The German manual is more
detailed and is explaining around a concept and sub-concepts. The Australian manual is more
basic, giving easy explanations and more step-by-step procedures, which are easy to understand
and follow when designing the system.
Chapter 2
PV Inverters
DC power cannot be fed directly into the utility grid. Power electronic circuits in the inverter are
therefore essential to produce the maximum amount of power (MPP). The DC power produced
in the PV array is converted to AC power with the required output voltage and frequency by
the inverter, making it ready to be used by electrical appliances or exported to the electrical
grid [5], [6]. This chapter presents the basic topology for a full-bridge voltage source inverter
and the dierent types of inverters available on the market.
2.1
The full-bridge voltage source inverter in Figure 2.1 consists of four switches and four freewheeling diodes. The four switches are turned ON and OFF so that only the diagonal switches
are on at each half-cycle (e.g. S1 and S4). The load is then alternately connected to the DC
source in each direction. The system operates with a time period between 0 and T . Between
0 and T /2 two switches are ON and V0 = +E, between T /2 and T the other two switches
6
Chapter 2 PV Inverters
frequency f = 1/T as shown in Figure 2.2. The frequency is controlled by the on- and o- rate
of the switches and gives a square wave voltage with E as its amplitude [5]. By turning the
switches ON and OFF several times in a cycle, it is possible to create an harmonic profile
similar to a sinusoidal.
v0
+E
T/2
-E
S1, S4
S2, S3
S1, S4
Figure 2.2: Output square waveform from the full-bridge voltage source inverter [5].
The switches in Figure 2.1 can be replaced by semiconductor switches as seen in Figure 2.4.
There are several types of semiconductor switches used in general electronic circuits, but the
majority of them are rarely used. Semiconductors mostly used in modern inverters are usually
MOSFETs, which are used at smaller power ratings, and IGBTs which are used at larger power
ratings [7], [8]. A semiconductor switch can be turned ON and OFF by using a comparator
that compares an input voltage to a reference signal and thereby turn the semiconductor switches
ON or OFF, depending on the results from the compared signals. Figure 2.3 illustrates two
comparators and how comparison of the input voltages determines if the IGBTs in Figure 2.4
are turned ON or OFF [9].
Pulse Width Modulation (PWM) is an efficient method to control the output of the inverter
and ensure stable RMS output value. By using PWM, the output voltage of the inverter can
be controlled by modifying the width of the pulses. A small control signal in direct proportion
with the pulses controls the width as a higher or lower control voltage increases or decreases the
width of the pulses, respectively. The control voltage can consist of a sinusoid of the desired
frequency and thereby produce a waveform where the average voltage varies sinusoidally. If the
control voltage to the IGBTs in Figure 2.4 is 0 V, then vu and vv are identical and vload is
0 V. By assuming there is a constant positive input voltage vin , which is equal to
1
2
of the peak
reference voltage vx and vy , then vload will be a train of pulses with the same width i.e. 50 %
Chapter 2 PV Inverters
Chapter 2 PV Inverters
duty cycle. If the input voltage vin is a sinusoidal voltage, the width of the train pulses will
vary sinusoidally with vin and produce an output voltage vload as seen in Figure 2.5 [9].
Figure 2.5: The output of the full-bridge voltage source inverter with a sinusoidal control
voltage applied. a) is the voltage input to the comparator, b) is the voltage pulse from T2 and
T3 in Figure 2.4, c) is the voltage pulse from T1 and T4 and d) is the output voltage pulse over
the load [9].
Figure 2.6 illustrates how the DC voltage (flat) runs through the full-bridge inverter and by
using PWM can achieve an approximate sine-wave output from pulse waves provided by the
switches [3]. As seen from 180 VAC raw, the PWM is a source of harmonic distortion, which
need to be filtered through an inductor. All of the harmonics is, however, not always filtered
out and a capacitor is often added (to create an LC-filter) to reduce the remaining harmonics
[8].
2.2
Inverter Transformers
Grid Connected inverters can be divided into three basic concepts and technologies: With a
low frequency (LF) transformer, high frequency (HF) transformer and transformer-less inverter
[10]. The LF and HF transformer electrically isolates the DC circuit from the grid (galvanic
Chapter 2 PV Inverters
10
isolation). Galvanic isolation is an expression for a topology in which the output circuit and the
input circuit are isolated or separated both physically and electrically, removing the connection
between output wires and input wires [11]. The advantages of using a LF or a HF transformer are
therefore the electrical isolation between the AC and DC side of the transformer and transformer
reliability because of few components. The HF transformer operates at higher frequencies and
is more compact and lighter than the LF transformer. It also has less transformer losses and
therefore a higher degree of efficiency when compared to LF transformer [10], [12].
There are a few inverters on the market with no transformer and they have a higher efficiency
than inverters with a transformer (no transformer losses). This does, however, lead to a disadvantage; it has no galvanic isolation. Some extra safety measures must therefore be added, like
a residual circuit breaker, for the inverter to be safe for use [10]. A DC-DC boost converter is
added in the front of the inverter to provide the DC voltage needed to provide a suitable AC sine
wave, like explained in Figure 2.5 [3]. When operating with a transformer-less inverter, there
is a risk of supplying small amounts of DC current to the utility grid. A utility transformer is
often installed within the switching sub-circuit of the inverter to prevent this and provide the
necessary required isolating characteristics [13], [14]. The description of a transformer-less inverter could therefore in practice be changed to a less-transformer (fewer transformers) inverter
[13].
2.3
Grid-connected Inverters
On the solar inverter market, manufacturers mainly produce two main types: Grid-connected
and Stand-Alone inverters. As shown in Figure 2.6, the grid-connected inverter uses a grid
sensing device and a digital signal processor (DSP) to ensure the inverter has a grid to connect
to and copy its voltage amplitude and frequency. This is to ensure the grid-connected inverter
produces exactly the same voltage and frequency as available on the grid, this is also known as
grid-compatible sine wave AC electricity [15]. If the grid sensing device in Figure 2.6 does
not find the voltage and frequency to copy, then switches in the inverter will disconnect it from
the grid.
Chapter 2 PV Inverters
11
Figure 2.6: The basic topology of a grid connected, single phase, transformer-based inverter
(at 60 Hz) [16].
2.3.1
MPPT
Grid-connected inverters normally use Maximum Power Point Tracking (MPPT) so the inverter
can operate near the maximum power point and keep the output efficiency as high as possible
[17]. The MPPT is an electrical system, which uses several dierent module electrical operating
points to optimise the module performance. The tracker searches for the arrays MPP at some
specific times of the day to ensure that the environmental conditions have not moved the MPP
[3]. The basic principle of the MPPT algorithms is the use of
to find the array output MPP. The algorithm measures both
P/ V = 0 (from P = I V )
P and
momentary operating region and is using this when increasing or decreasing Vref erence . By doing
this, the MPPT algorithm ensures that the PV system operates by the MPP of the array at all
time [17]. This method can, however, not track the MPP under rapid and instant changes in
irradiance conditions. A more complex method called the incremental conductance method is
an alternative [17], but will only be mentioned in this thesis.
2.3.2
Dierent types of inverters are developed for the market according to their power ratings and
depending on how large PV array the customer wants to connect to a single inverter. This
varies from small modular inverters (micro inverters in AC PV modules) to large PV-inverters
Chapter 2 PV Inverters
12
generating 100 kW or more. An overview of the dierent inverter options is illustrated in Figure
2.7.
Central Inverters
Central inverters are normally used with a power rating from 30 kW to 1 MW in large gridconnected PV systems [3]. The largest solar power plant in California is located on Harper Dry
Lake in the Mojave Desert with a power rating of 80 MW AC. A new solar power plant is under
construction in Colorado Desert with an estimated power rating of 550 MW AC [18]. These
are all using large central inverters with a high power rating. This system has the advantage
of having all the components located at one location for maintenance, good inverter efficiency
and lower cost per watt [3].
String Inverters
A string inverter is the most common inverter used in small grid-connected PV-systems. Some
string inverters can serve up to two or three strings of PV modules, with an inverter power rating
from 1 kW to 12 kW. However, with a single MPPT it is recommended to use a multi-string
inverter for arrays over 5 kW [3]. The downside of having multiple strings in one string inverter
is that only one MPPT is installed in each inverter. If one string is shaded then this will aect
the output of the entire array, instead of just one string. Individual string inverters for each
string could be an alternative but would also be more costly [3]. As illustrated in Figure 3.3,
PV arrays based on string inverters are more flexible than PV arrays based on central inverters
since they are both easier and cheaper to extend.
Chapter 2 PV Inverters
13
Figure 2.7: An overview of the dierent inverter options: a) Central inverter, b) String
inverter, c) Multi string inverter, d) Modular inverter [7].
Chapter 3
Grid Connection
To ensure from whom and how power is being fed into the grid, there exists regulations to
control this. These regulations made by NVE and by the government makes sure that the grid
is balanced at all times and that no other than utility companies and plus customers can
feed power into a grid. To get permission to become a plus customer an application to Agder
Energi needs to be fulfilled according to the given regulations and standards.
3.1
The Norwegian water- and energy directorate (NVE) defines a plus customer as an electrical
consumer with a yearly production that usually does not exceed its own consumption, but at
some times of the day has an excess production of electrical power that is fed into the grid [20].
For the implementation of such a small scale power plant to be possible, NVE gives exemption
from some of the requirements larger power utility companies has to sign, such as the balance
contract, trading permit and the tari contract for feeding electricity to the grid. The purpose
of this exemption is to make it easier for plus customers to feed and sell their excess power to the
utility company. The plus customer still has to sign a grid connection- and grid lease contract
with the local utility company to ensure the quality of the power being sold. For the power to
be sold, NVE requires the plus customer to install a two-way smart meter in the switchboard
[20]. This is however not required if the electrical power produced is intended for private use
only, which is the case in the current project where the maximum PV system output will never
be greater then the campus consumption [21].
14
3.2
15
Grounding Systems
The local grid at UiA Grimstad is using a 400 V three-phase TN-S system. A TN-S configuration
is the way a system is grounded and means that the neutral and protective earth are separated.
The grid is then grounded at the transformer or generator end (T or Terra is French, meaning
Earth) while the neutral conductor at the end user is grounded [22]. A TN-S system is illustrated
in Figure 3.1.
Inverters with a relatively low power, normally String Inverters, operate as single-phase inverters. Whether or not to use a single-phase- or a three-phase- inverter depends on the size of
the system, a decision need to be made before further planning of the PV system is done. As
a general rule, PV-systems lower than 5 kWp are built as a single-phase system and larger
systems than 5 kWp are built as three-phase systems [4]. Figure 3.2 illustrates the dierences
when connecting an inverter to the two phase-systems, the inverter in single-phase systems is
connected between one phase at the grid and the neutral, for three-phase systems the inverter
is connected to each of the three phases of the grid as well as the neutral [4].
It is becoming more common to use multiple single-phase inverters in a PV system [4]. These
can be connected to one or three phases, depending of the inverter power. A system cannot feed
more than 15 kVA into one phase; the inverters should then be equally divided so that the load
on each phase is equal. Figure 3.3 shows a typical configuration where multiple single-phase
grid connected inverters are used in an array and connected to one phase via a circuit breaker
16
(Q4). The design can be used for all kinds of buildings and can be duplicated as many times as
necessary [23].
Figure 3.2: Configuration of a single-phase inverter connected to one phase, and a three-phase
inverter connected to all three phases [4].
Figure 3.3: Illustrating a system with multiple single phase inverters connected to the grid
via switch Q4 [23].
Whether or not to use a single- or three-phase system is also determined by what the manufacturer can provide. A look at three inverter manufacturers, Eltek, Danfoss and SMA shows that
the smallest grid connected inverter producing 400 V at three-phases was the SMA Tripower
5000 TL-20 producing 5 kW. The smallest three-phase inverter manufactured by Danfoss is the
TLX- and FLX-series producing 6 kW, and the smallest three-phase inverter manufactured by
Eltek was the Theia TL producing 13 kW. Neither of the three manufacturers has a 400 V
three-phase inverter with a transformer, which is recommended by Agder Energi.
17
The element of uncertainty by installing a single-phase system is the phase balancing. All threepower lines can carry the rated power in both directions. A large voltage source (inverter) on
one phase would lead to a change in potential between the three phases. The phases would then
be unbalanced and higher currents would flow through the neutral leader. Agder Energi does
not, at this time, have regulations for how much power they allow to be fed into one phase. The
regulation is under development [21], but Hafslund has set their upper limit for single-phase
production to 15 kVA so it can be assumed that this number also applies for Agder Energi [24].
This assumption will result in a maximum allowed phase-current of 65.22 A for a 230 V phase.
3.3
Islanding
Islanding is a critical phenomenon and is at the time not wanted on the utility grid. It occurs
when the PV system keeps on feeding electrical power to a utility grid that is isolated from
its utility voltage source. The utility grid can be isolated when a grid and a PV-inverter are
separated by accident, on purpose or by damage [7]. This is usually divided between wantedand unwanted islanding, where wanted islanding occurs when renewable energy sources are
connected to make a micro grid as shown in Figure 3.4. During islanding, this micro grid is
isolated from the utility grid and the renewable sources are used to power it. This micro grid is
then called an island [25].
Unwanted islanding can present a high risk for utility workers maintaining the grid as well
as causing unwanted incidents on the utility grid. Utility workers arriving to do scheduled
maintenance work somewhere on the isolated island might think that it is not energised since
the switch (showed as CB1 in Figure 3.4) has to be manually disconnected. Just a few tens
of seconds would be lethal if utility workers had commenced maintenance work on the grid
immediately after the switch (CB1) was disconnected [26]. Even though the grid is manually
disconnected from the micro grid, local production units can produce power a few more seconds
until the grid-sensing device cannot find the grid, or until the inverter anti islanding device kicks
in. The AC switches (CB1 and CB2) should therefore have an indicator on each side, indicating
if AC power is present or not, and to clarify this by an alarm or a light.
Figure 3.4 is an example where the local production unit is the micro grid. PCC is the Point
of Common Connection where the micro grid is connected to the main grid, CB1 is the circuit
breaker between the main grid and the micro grid and CB2 is the AC circuit breaker between
18
Figure 3.4: Solar station disconnected from the utility grid, and operating as an isolated
island [27].
the grid connected inverter and the grid it is connected to. In situations where the PV system is
placed on the roof, the PCC is situated in the house switchboard where the power Pload + jQload
is supplied to the house and a
grid. CB1 is still the circuit breaker connecting or disconnecting the main grid to the micro
grid. In cases where the grid needs to be disconnected for maintenance, the switch CB2 must be
disconnected prior to disconnecting CB1, and connected after CB1 to avoid islanding. It must,
therefore, be possible to remotely control the inverter so this can be done directly by the utility
company. Many inverters now have communication terminals e.g. CAN interface (Controller
Area Network) that are intended for smart grids or to communicate between a master and
several slave inverters. This allows many local PV- or Wind- production units to supply a
micro grid and be remotely operated by the local utility company..
3.4
Agder Energi has the sole responsibility for operating, maintaining and ensuring the power quality in the grid. The documentation needed for the application is found on www.aenett.no. The
application process can then be divided into five steps found when studying the documentation
from Agder Energi. Further discussions of these 5 steps will be done in Chapter 5.7.1.
1. Study the technical guidelines in Appendix J for feeding power into the grid. The chosen
inverter must fulfil the restrictions given in Chapter 5.2 and 5.3 for it to be approved by
Agder Energi.
19
2. The contact person for the PV system must complete an application form (Appendix
B) made by Rasjonell Elektrisk Nettvirksomhet (REN). For UiA this person is Kristen
Leifsen, building manager at UiA campus Grimstad. Required information to fill in is the
inverter name and type, the installed power, estimated summer- and winter-production in
GWh and development cost of the plant. Agder Energi estimated the time elapsed from
the application form was received to a permission was granted to be a few days (in reality
it took longer time).
3. After the permission is granted Agder Energi needs to start working on an agreement
between them and the customer (UiA) for how the power should be sold regarding price
and quantity. In addition, they wanted a preliminary single line diagram (Appendix G)
for how the system would be built regarding fuses, circuit breakers, cables, grounding
and galvanic isolation. The agreement between the customer and Agder Energi, given in
Appendix L, explaining the dierent parties rights, dierent conditions for grid connection
and stating that the customer has read all of the additional documents then need to be
signed and sent to Agder Energi.
4. A site Acceptance Test must be made of the grid before a PV-system is connected.
5. A documentation form (Appendix K) has to be signed by the installing electrician explaining the protection installed and installations made to ensure that the requirements
in FoL are followed (frequency, step voltage, unbalanced loads, flicker, harmonic).
Chapter 4
4.1
The sun is by far the largest energy provider and without it the earth would be yet another
dead planet in the universe. Since all energy (with the exception of geothermal energy and
nuclear) on earth originally derives from the sun, it should come as no surprise that no other
energy source could match the suns energy potential. The energy in the sunlight reaching
the earths surface is approximately 10,000 times the worlds energy requirements, and only
0.01 % of this energy is needed to cover the humans energy need [4]. Solar radiation on earth
is influenced by several factors; earths atmosphere, shadowing, latitude, season and time of the
day, etc. The yearly average extra-terrestrial solar irradiance is approximately G0 = 1, 367
W
.
m2
This irradiance is reduced in the atmosphere by scattering caused by aerosols, particles and
20
21
absorption by atmospheric gases. This reduces the incoming sunlight by around 30 % [28].
AM =
1
sin
(4.1)
Equation (4.1) describes the air mass solar light must pass through in order to reach earths
surface. In other words, cosecant is the ratio between the suns elevation path and zenith. If
the sun is directly overhead the solar light must travel through less air mass then if the sun
had a lower zenith angle
and depends on the geographic location and time of day. The irradiance on a sunny winter day
at noon in Grimstad could be 600
W
.
m2
W
m2
W
m2
seen in Figure 4.1. A video showing the sky conditions needed to create over-irradiance is
TM
available on Youtube
1.6
PV Cell
Pyranometer
1.4
1.2
1
0.8
373
374
375
376
Time [ min ]
377
than further
south in Europe. The solar irradiation in Norway is seasonal and varies substantially throughout
the year. Norways extensive length yields large dierences in annual solar irradiation, from
700
kW h
m2
kW h
m2
overview of solar irradiation in five Norwegian cities and Hamburg located in northern Germany.
The PV modules are mounted at the optimum tilt angle and orientation (south) as calculated
from PVGIS, verifying that Kristiansand in the south has the highest potential for using solar
power generation in Norway. It is no surprise that Hamburg receives a higher irradiation than
the Norwegian cities since mapping of solar irradiation in Norway is poor when compared to
1
h
Solar irradiation is the amount of solar energy falling on unit area over a stated time interval [ W
].
m2
22
Central Europe, thus the solar irradiation could probably be higher than the estimations given
in PVGIS. More on this in Section 4.1.2. These annual irradiations are significantly aected by
the cloudy weather, which is common in these regions. According to [32], in 2011 nearly half of
the irradiation measured in Grimstad, 1200
kW h
m2
of 2130
kW h
,
m2
utilised in the PV modules. Research from Northern Research Institute (NORUT) indicates
that the solar power in certain areas in the nordic countries could reach similar energy harvest
as Germany [33]. However, in order to accomplish this, the PV modules need to be installed at
optimum angle and tracking devices would be needed to ensure maximum incoming irradiance.
Adding a tracking device to the PV system would at the moment not be beneficial due to high
costs in the small immature PV market in Norway today.
Table 4.1: Average sum of global irradiation per square metre received by the modules of the
h
given system ( kW
m2 ). Mounting position towards south, with optimum inclination. Radiation
database: Classic PVGIS (based on ground measurements) [34].
Month
Inclination
January
February
March
April
May
June
July
August
September
October
November
December
Monthly average
4.1.1
Oslo
40
16.9
39.4
75.7
115
150
156
151
122
89.3
50.7
23.0
13.6
83.5
Bergen
33
5.49
27.6
72.2
115
144
158
144
111
76.7
36.7
7.92
3.13
75.1
Tr.heim
43
15.1
47.2
94.2
128
153
155
147
121
81.9
49.1
21.5
8.16
85.1
Kr.sand
38
22.0
43.5
81.7
120
155
155
154
127
93.4
55.7
26.5
15.5
87.4
Troms
45
0
25.4
86.6
140
144
144
138
106
69.1
27.0
0.548
0
73.4
Hamburg
36
27.5
51.2
79.9
125
159
140
151
141
101
69.8
36.2
20.3
91.8
In order to calculate solar irradiance on inclined surfaces the solar radiation is decomposed into
direct beam, sky diuse and radiation reflected from the ground. Depending on the ground, a
coefficient (albedo) is used to represent the diuse reflections from the ground. The estimated
irradiance on the inclined plane can be calculated by using Equation (4.2). The global tilted
irradiance (Gt ) describes how much solar power per unit area (W/m2 ) the solar modules are
exposed to.
Gt = Gbt + Gdt + Grt
(4.2)
23
Surface
Desert
Dry sand
Soils
Water (high sun)
Water (low sun)
Fresh snow
Melting snow
Albedo
0.25 - 0.30
0.2 - 0.4
0.05 - 0.20
0.05 - 0.10
0.5 - 0.8
0.75 - 0.95
0.35
surfaces. In Norway, reflections from the many lakes and ponds may give an increased solar
energy generation. As seen in Figure 4.2 from the installation site the sea at the distant horizon
could reflect and contribute with additional irradiation to the PV system. Although the water
has low albedo values ranging from 0.05 to 0.10, it can probably be assumed that the overall
size of the reflecting area would compensate for this. Further research and measurements are
required on the site to verify and determine how much solar irradiation sea reflections would
contribute with. During wintertime, reflections from snow and ice must also be taken into
consideration as contributors to the global irradiation.
4.1.2
Solar radiation can be measured by ground sensors or satellite. Typical pyranometers and PV
based sensors are used as ground sensors, but there are also other alternatives such as photodiodes. A pyranometer is built up by one or two hemispherical glass domes with a thermopilesensing element in the centre, which is encapsulated by the instrument body. The solar radiation
is absorbed by a black surface on the thermopile sensor and compared with a white surface (white
24
Figure 4.2: Southern view at the installation site with the sea in the horizon [19].
metallic housing) that reflects the solar radiation. The temperature dierence is measured and
the thermopile sensor converts this thermal energy into voltage proportional to the temperature
dierence. The glass dome restricts the field of view to 180 while operating on a spectral range
from 300 to 3,000 nm. More expensive pyranometers are equipped with two glass domes to
improve thermal insulation and reduce the directional error.
A PV sensor consists of a solar cell and is usually made of crystalline silicon. The sensor operates
by delivering a voltage proportional to the measured irradiance. The main benefits when comparing a PV sensor to a pyranometer are low cost, similar reflection losses as in PV modules and
fast response (A pyranometer needs longer measurement times due to thermal transients). Disadvantages are limited spectral sensitivity and higher inaccuracy than a pyranometer [4]. The
spectral range for a c-Si PV based irradiance sensor is approximately from 300 to 1,100 nm.
Images from geostationary satellites can also be used to derive global irradiance on earths
surface. This is done by using the software Heliosat-II, which converts images from satellites
into maps of global irradiation at ground level. This method is described in detail in [36].
The newest operational satellites have a resolution of 1 km x 1 km near the equator below the
satellite. This resolution is reduced at high latitudes due to low viewing angle, the satellites are
therefore not suitable to map the irradiance north of 58 in Norway [37].
PVGIS
One of the most important factors when deciding on installing a PV system, is to have the
available radiation data on the possible installation site. A popular tool for estimating the
25
potential solar radiation is PVGIS developed by the JRC of the European Commission. Solar
irradiation data can from any location in Europe be obtained by using this web-based application. PVGIS utilise interpolated ground station measurement data from the period 1981 to
1990 (PVGIS-3 database), satellite images from 1998 to 2005 and from June 2006 to December
2011 (PVGIS-CMSAF database). Satellite data is used south of 58 N [34], [38] and therefore,
only the most southern parts of Norway have irradiance data from satellite images. The ground
measurements are retrieved from 566 meteorological stations around Europe. In Norway, data
is used from one meteorological station located in Bergen. The measured data from this station
is interpolated with other stations in Europe. According to [34], the old database with interpolated ground measurements (PVGIS-3) estimates lower irradiance than the new database, which
are based on satellite images. Comparison with newer ground station measurements show that
the new database has a higher accuracy than the old database. This is likely caused by wrong
measurements in the old database due to shading, frost, snow, dirt, etc. Another factor is the
increased solar radiation over Europe in the last 30 years, contributing to higher irradiance in
the new database. In Norway, it would therefore be sensible to compare the PVGIS data with
local measurements. The company Kjeller Vindteknikk has established a database with data
from 68 stations measuring solar irradiance in Norway and local data from this database could
then be used as a validation when working with PVGIS in Norway [37].
Local Irradiance
The Photovoltaic Lab at the University of Agder has been logging irradiance data since 2011.
The sensors are mounted 7 east from south and with 39 1 tilt angle from the horizontal.
The irradiance is measured with a Kipp & Zonen CMP 3 Pyranometer and the PV based
sensor 80SPC from Soldata. The results are presented in Table 4.3 together with a long term
prediction of PVGIS. Due to little importance in PV power generation, values below 50
W
m2
from the local measurements performed in 2011-2013 are discarded. The monthly average
irradiation for each year is slightly higher than given by the PVGIS prediction. The highest
monthly average irradiation was 108.8
sunny month with 201.69
kW h
m2
kW h
.
m2
kW h
m2
which is 16.2 % higher than what PVGIS estimates. This indicates higher irradiance in
Grimstad than given by the long-term average in PVGIS. Only the old database was available
for Grimstad and some of the reasons explaining the lower irradiance from PVGIS is stated in
section 4.1.2. Another possibility could also be as mentioned in section 4.1.1, the albedo eect,
26
Table 4.3: Average monthly global irradiation per square metre received by the modules of a
h
given system ( kW
m2 ). Mounting position is 7 east from south, with 39 1 tilt angle from the
horizontal. Location: Agder Photovoltaic Lab, University of Agder [39].
Month
January
February
March
April
May
June
July
August
September
October
November
December
Monthly average
2011
48.1
28.0
118.0
166.0
156.0
169
149.0
145.0
101.0
69.4
26.8
23.2
100.0
2012
22.14
59.77
128.29
115.59
188.52
153.63
169.43
158.01
106.56
72.92
26.35
11.18
101.0
2013
31.32
63.77
143.31
160.21
136.78
164.98
201.69
161.13
109.78
60.21
51.87
20.58
108.8
PVGIS
21.0
43.3
80.0
119.0
152.0
155.0
153.0
127.0
92.7
55.1
26.0
15.4
86.6
which would contribute to the globally tilted irradiation due to reflections from the sea, snow
and ice. The measurements obtained in Grimstad look promising and show that PV systems in
Southern Norway could be a viable energy source with government support, or even without if
the current PV prices decrease. However, the information presented here is only an indication,
as no final conclusions can be made until there is enough data (7 to 10 years [40]) to support
these findings.
4.2
The available solar irradiance and the ambient temperature aect the power output. With a
nearly linear relation, the current is highly dependent of the solar irradiance incident on the
solar cells. If the light intensity is decreased by 50 %, the same reduction can be measured in
the current. The short-circuit current also depends on the spectral composition of sunlight. The
degree of the dependence varies amongst dierent PV technologies. The voltage is not aected
to the same extent as the current, but there will be a slight increase with higher irradiance [3].
Temperature is the second most important parameter after irradiance aecting the electricity
output. If the temperature increases above 25 C the voltage will decrease and thus reducing
the power output. One of the benefits in Northern Europe is the cold climate reducing the
ambient temperature of the solar cells. When temperatures are lower than 25 C the power
output increases. This is shown in Figure 4.3(a) which illustrates how the voltage decreases
27
Figure 4.3: Eects of temperature and irradiance on the solar cell performance.
when temperature is rising. Figure 4.3(b) then shows an I-V curve at a constant temperature,
indicating how the current and voltage are dependent of the irradiance. Open-circuit voltage
VOC , short circuit current ISC , efficiency and fill factor are all temperature dependant. Due to
low solar cell efficiency when converting irradiance to electrical energy, typical 9-12 % of the
remaining irradiance is either reflected or converted to heat. The generated heat increases the
temperature of the PV modules resulting in decreased efficiency.
4.3
Ambient Temperature
The PV array sizing involves the maximum and minimum operating temperatures of the PV
cells with the intent of determining the best and worst VOC and VM P . This is dependent of the
ambient temperature at the location of the PV system, a temperature that is measured in a
shadow so that direct solar radiation does not aect the result. Temperature measurements at
UiA Grimstad Campus has only been operating since late 2010 which is not a sufficient number of
measurements to be used. The Bioforsk AgroMetBase temperature data for Landvik, a location
5 km from UiA Grimstad, is considered sufficient to use in this thesis. This database can
provide data for wind, temperature, air pressure and density in maximum, minimum or average
values for every hour, day or month of the year. Maximum and minimum air temperatures at
two-metre altitude are used as a starting point to find the operating temperature of the solar
cells [43]. The maximum values from Landvik are only registered since 1 January 2006, which
is not considered sufficient for good measurements but is still used to give a general idea of
the temperatures in the area. From temperature data and Figure 4.4, the maximum ambient
air temperature at two-metre altitude for Landvik was 29.4 C measured on 11 June 2007.
28
Figure 4.4: Maximum ambient temperature of 29.4 C measured on 11 June 2007 at two
metre altitude for Landvik.
The data for the minimum ambient air temperature started logging on 10 April 1987. It is
considered that 26 years are a sufficient amount of data to paint a good picture of the minimum
temperature in the area. It can be seen from Figure 4.5 that the lowest ambient air temperature
at two-meter altitude for Landvik was
29
(4.3)
TA is the ambient air temperature, G is the in-plane irradiance and Ct is a constant dependent
of the module mounting and ventilation. It is assumed that the ratio between T and Ta is a
linear function of the irradiance (G) on a module [44]. It is also assumed that the maximum
steady irradiance is approximately 1200 W/m2 during summer and 800 W/m2 during winter
in Grimstad [34]. When sizing the array, one need to know the highest VOC and lowest VM P .
For best and worst case scenario, we need to find the highest and lowest operating temperature
which is found by choosing Ct = 0.035 C W/m2
during winter
[34]. The maximum and minimum module temperature at peak irradiance is calculated to be:
Tmax = (29.4 + 5) C + 0.035 C W/m2
(4.4)
5) C + 0.022 C W/m2
800 W/m2 =
(4.5)
Tmin = ( 22.8
4.4
10.2 C
The voltage drop equals the total resistance times current. The total resistance is dependent of
the cross-sectional area of the conductor, the wire length and resistivity. The voltage drop is a
loss of voltage in the cable due to resistance and causes a loss in power, expressed as P = I 2 R.
The Australian standard AS/NZS5033:2005 is specifying the voltage drop from array to energy
meter to be less than 5 %, meaning that the PV string voltage drop plus array voltage drop
plus AC cable voltage drop is less than 5 % [32]. This will be explained more in detail when
sizing the cables in Section 5.4.2. The voltage drop in the DC cable can be calculated using
Equation (4.6).
VDrop DC =
(4.6)
LDC cable is the length (m) of the DC cable one way, assuming symmetry in the cable both
ways, IDC is the DC current (CCC) in the cable at worst case scenario, is the wire resistivity
mm2 /m ) and ADC
cable
is the cross section of the DC conductor mm2 [32]. For the sum
30
of the voltage drop in the string- and array cable to be less than 5 %, it is assumed that a
maximum voltage drop of 2 % is used in the string (module cables) and the array cables.
4.5
Wind load
UiA- Grimstad is located on a hilltop 45 metres above sea level, with a building height of about
15 metres the PV array is located at approximately 60 metres above sea level [19]. This hilltop is
exposed to a large variation of wind forces from both sides of the array, creating a windward side
of the array with a high-pressure area and a leeward side with a low-pressure area. The wind
suction on the leeward side could be strong enough to break the glass or tear the PV modules
o its mounting rack if high enough uplift forces are created [4]. The dierent scenarios are
illustrated in Figure 4.6. As illustrated in Figure 4.6(a), the windward will produce forces and
Windward
Leeward
Windward
Leeward
stresses on both the PV modules and the array structure in addition to a vertical force eect
and a fairly high drag. Most of the loads would, however, occur in the vertical direction [46]
and eect the supporting beam and PV modules. Figure 4.6(b) illustrates windward forces
from the rear working in a vertical direction, lifting the array upwards and inflicting tension
to the support beams [46] and PV modules. Forces on the array will be divided between the
modules, creating stresses on both supporting beams and PV modules. Because the array is
located on the roof of the building, where the wind loads are most intense, building engineers
who did the wind load calculations for the university followed the Norwegian standard NS
3479; they did, however, state that a wind load of 1.5 kN/m2 could be used instead [47]. PV
modules used in the array has an area of 0.84 m2 , they would have to withstand a load of
1.5 kN/m2 0.84 m2 = 1.26 kN each. As mentioned earlier, wind forces on the array will in
general aect the PV modules and the supporting beams. The array is connected to the roof by
31
steel beams anchored in concrete. Very high forces must therefore be present for the array to
lift o. Stainless steel bolts connecting the array to the supporting beam will secure the array.
However, one risk is that the PV modules could take o as in Figure 4.7. This will be prevented
by fastening the PV modules with four to six nuts and bolts depending on the distance between
the beams. This produces another risk; if the module frame is securely fastened the wind could
blow through the PV modules and destroy them by creating a large hole or by bending the
surfaces so the glass breaks. As seen in Appendix M there has not been done any testing of
how much wind load one PV module can resist, and we know that the maximum wind load
on each PV module can be as much as 1.26 kN. Depending on how they are mounted, modern
PV modules from SunTech are tested for extended wind and snow loads, and sell PV modules
certified to withstand wind loads between 2400
rear) and snow loads between 2400
The expected maximum wind load on the roof is much lower than the minimum wind load
sustainable by modern PV modules. By comparing the minimum wind load 2400 N/m2 with
NAPS (NESTE) PV modules they would be able to withstand 2400 N/m2 0.84 m2 = 2 kN .
Assuming the NAPS modules can withstand the same as modern PV modules, they would be
more than able to withstand the expected wind load of 1.26 kN.
Figure 4.7: Illustrates PV modules that got destroyed during a storm [49].
4.6
Temporary Shading
On the installation site, typical temporary shading would come from snow, bird droppings and
the surroundings. It can in general be assumed a power loss between 2 % to 5 % due to soiling,
but according to [4], this is accepted. The university roof has become a nesting ground for
seagulls in the area, so bird droppings on the PV modules like in Figure 4.9 is expected to
happen regularly. The PV system is to be mounted on the roof where a fence (roof barrier)
32
is installed on the roof edges. At a certain time of the day, this roof barrier cast a temporary
shadow on the lower PV modules. In Figure 4.8 both shadow from the snow and the fence is
projected on the PV modules. Due to the set-up of the rack it is not possible to avoid the
temporary shading from the fence. There are many factors contributing on how large eect the
shading would have upon a PV system. First of all, it must be determined how many modules
that could be shaded simultaneously. By studying Figure 4.8 it can be seen that only the
modules on the two lower rows are partly shaded. The array configuration for the PV system
in this thesis can be seen in Figure 5.4. Due to limited space on the rack, other test modules
and the placement of sensors makes it difficult to completely avoid the temporary shading from
the fence. When comparing Figure 4.8 with the module configuration in Figure 5.4 it can be
seen that modules 18 to 23 and 1 to 4 is most likely to be aected by temporary shading.
Figure 4.8: Partial shading from the fence on previous installed modules. Photo was taken
9:30 on January 2011 [19].
Figure 4.9: Bird droppings are expected to occur regularly at the site [19].
4.7
33
Corrosion
PV modules are designed to operate for more than 25 years in the field, and it is important
that the rest of the system target a similar lifespan. Mechanical fittings used on rack mounted
PV systems are exposed to dierent weather conditions. Low steel quality and combinations
with dierent metals in the fittings increases the risk of corrosion [4]. The location plays a large
role on how exposed a PV system is to corrosion. Salt from the highways used for de-icing
during the Norwegian winter, as well as from the sea and the wet climate are factors needing
consideration when designing a PV system. There are several types of corrosion, such as pitting
corrosion, stress corrosion cracking, galvanic corrosion, crevice corrosion and erosion corrosion
[50]. According to [50], galvanic corrosion is the most relevant corrosion eect as it has often
been seen on PV plants. Galvanic corrosion is an electrochemical reaction happening between
two pieces of dierent metallic elements. In order for galvanic corrosion to occur, one element
needs to function as an anode and the other one as a cathode, in addition there needs to be an
electrolyte and conductivity between the elements [51]. As seen in Table 4.4, elements possess
dierent electrode potential and give an overview on the elements most likely to behave like an
anode or cathode when combined. For example by combining zinc and nickel, the zinc would
behave like an anode and corrode respectively. Metals in a PV system, such as aluminium,
Table 4.4: Galvanic series, showing metallic elements and its corresponding electrode potential
[50].
Element
Magnesium
Beryllium
Aluminium
Zinc
Chromium
Iron
Cadmium
Nickel
Tin
Hydrogen
Copper
Silver
Platinum
Gold
stainless steel and copper make up the anode and the cathode. The electrolyte is made of water
from rain, moisture, sea, etc. Fixing points is most likely to be aected by corrosion, but can
be prevented by following this guideline from [51]:
34
When mixing metals, choose metals close together in the galvanic series.
Minimise contact area between dissimilar metals.
When mixing metals, small parts (fasteners) and critical components should be of a more
noble element.
Avoid threaded joints for elements far away in the galvanic series.
Apply coating or insulate dissimilar metals.
Modules already installed on the planned location for this project had some cosmetic corrosion
between the washers and the frame of the module, but there were no signs of any physical damage. The frames of the solar modules and the support beams are made of anodized aluminium.
By anodizing the aluminium, a thick oxide layer is made to protect it from galvanic corrosion.
The bolts, nuts and washers are made of austenitic stainless steel with Type 316 material,
which is an alloy of iron, chromium and nickel. This is marine grade stainless steel and has an
increased corrosion resistance compared to other types of stainless steel, such as Type 304. The
chromium reacts with the atmosphere and form layers of chromium oxide, which protects the
steel underneath [51], [52].
Due to the precautions of choosing materials with a protective oxide layer it is unlikely that
any galvanic corrosion should develop on the frames of the modules, fasteners or the support
beams.
4.8
Grounding
The reasons for connecting the module frames (exposed conductive parts) between the PV
system and earth are apparent due to: Bonding in order to avoid floating potentials in the
PV system; provide a path for fault current; protection from lightning [3]. Bonding is used
to prevent dierent voltage potential to ground and reduce the risk of electrical shock due to
module encapsulation or isolation breakdown. The risk of the array together with the beam
support being an extraneous-conductive part is highly unlikely. The PV system will be installed
on the roof where there is no direct contact with the earth making it unlikely for the frame of the
array together with the beam support to introduce a potential into the installation. However, in
this project it is decided to equipotential the beam support and the frames of the PV modules
to the fence. This introduces the same potential to all the extraneous-conductive parts on the
roof and prevents the hazard of electrical shock. The DC side may be ungrounded or grounded
35
either on the negative or the positive terminal, but grounding on the DC side can only be done
if there is a simple separation (like a transformer) between the DC side and the AC side [53].
The grounding topology of a PVGCS can be seen in Figure 4.10. The NEK 400:2010 preferably
recommends the DC side to be constructed with Class II equipment (double insulation) or similar
insulation [55]. If the installation meets these requirements, it is not necessary to ground on
the DC side [56]. However, grounding is required on the AC side due to the inverter being a
Class I equipment. In this project, it is decided to follow the guidelines of NEK 400:2010. The
first topology, ungrounded, seen in Figure 4.10(a) is, therefore, applied to the DC side of the
PV system. The PV array is one of the highest points on the roof, including a large metal
building and a high tower. It is believed that the metal building is grounded by obvious reasons
and a grounded socket which is connected to the metal building is located directly under the
array. The university building is, however, not optimised for grounding PV systems. Grounded
objects are attracting lightning and it is not desirable to lead high voltages caused by lightning
through the building. It is therefore decided not to ground the PV system on the DC side but
rather ground it on the AC side, making sure that NEK 400:2010 is fulfilled.
4.9
36
Fire Hazard
As long as electrical appliances are involved, there is always a risk of fire and PV systems are no
exceptions. PV systems do not just present danger to buildings, but also fire-fighters or building
occupiers accidentally coming into contact with a live wire [57]. The risk of a building being
damaged because of a PV fire is estimated by [58] to be 30 10
risk is based on 30 annual fires for the 1,000,000 PV systems installed in Germany. The risk is,
however, higher for PV generators integrated into the roof (BIPV). These systems need extra
precautions and risk awareness when being installed. Stando PV generators are mounted on
hard roof that acts as a shield between the building and a PV fire. These stando systems also
have the advantage of being cooled down from the wind etc, because of the distance between the
PV module and the hard roof [58]. Between 1995 and 2012, some 180 fires were caused by PV
systems. A study of Figure 4.11 reveals that most of these fires originated in the DC section of
the PV system. Most of these fires did, however, originate because of a lack of maintenance and
bad practices during installation. Based on Figure 4.11(a) extra care must be taken regarding
components with highest counts when installing the PV system, these must also be inspected
annually with thermal imaging camera to ensure that no faults are building up with time [58].
Figure 4.11: Fire causes in a PVGCS indicating the component and section where most fires
start [58].
Chapter 5
5.1
Inverter Selection
The type of inverter depends on the size of the solar power plant. It is given that the inverter
should have a size of approximately 3.5 kW. This puts some limitations for the inverter type.
The two inverters most suitable for this PV array are string inverter and multi string inverter.
A multi string inverter would have been a better option if the customer thought about installing
multiple strings and maybe expand the solar power plant in the future. This is, however, not
an issue at the current location with a total area of 27 m2 to install a PV array, it would not be
very likely to expand the power plant or add more than two strings. The price is also increasing
when a multi string inverter is chosen. Even though a string inverter only has one MPPT it is
considered the best option for this small system. Most likely a single string will be used, but it
has the possibility of adding two more strings if necessary (Appendix A).
One of the criteria for choosing inverter is that a local supplier is used. A search for inverters
showed that the relevant inverters were not found close to Grimstad, but distributed over Norway
37
38
and Denmark. The design criteria for the inverter size were chosen due to the intention of a
long-term investment for the PV lab at UiA-Grimstad. Our supervisor wanted the possibility
of connecting modern PV modules with higher current and voltage ratings at a later time [19].
Inverter Types
PVS300-TL is a solar string inverter produced by ABB. This is a single-phase inverter with
integrated performance data monitoring and IP55 protection level (dust- and water-proof). It
is transformer-less which makes it smaller and lighter than other inverters. The monitoring
graphical display is detachable with its own keypad and an EIA-485 interface and Ethernet
monitoring protocol. Embedded in the inverter is a DC power switch, DC string fuses, reverse
polarity protection, RCD (residual current detection) and grid monitoring with anti islanding
[59]. A price quotation is given from ABB for the PVS300-TL to be 12800 NOK plus taxes and
transportation costs [60].
Getek in Trondheim distributes the SMA sunny boy S3800. The SMA S3800 has an integrated
DC circuit breaker and a maximum input of three strings, designed with a current limit of
16 A. It uses a LF transformer, which provides galvanic isolation, but it does not provide any
more information about data logging other than Bluetooth and RS485 interface is optional. Of
protective devices, it provides DC reverse polarity protection, AC short-circuit current capability
and an optional arc-fault circuit interrupter [61]. However, the datasheet is inadequate when
compared with datasheets for the other inverters [62]. It was not possible to get a price quotation
for the SMA sunny boy S3800 from Getek.
Eltek in Drammen is supplying the Theia HE-t 3.8. Theia HE-t has a maximum efficiency of
97.3 %, and high efficiency at low irradiation. It is manufactured after the IP65 protection
level, meaning that it is dust- and water-proof for installation outdoors or indoors. The inverter
uses one MPPT and has three string inputs. The input features reverse polarity protection,
ground fault monitoring, optional DC switch and DC string fuses and it is field configurable
for ungrounded, positive or negative grounding [63]. Appendix I shows the logging capabilities
of the inverter which include input- and output current, voltage and power as well as energy
and power produced daily, monthly, yearly and in total. Communication with Eltek explains
that Theia HE-t uses two separate log types; an energy log with 365 daily sums and up to
365 monthly sums and a data log containing data from every 15 minutes while the inverter
is operating. The data log is easily downloadable from the inverter on a PC via the built-in
39
Ethernet interface using a PC-tool provided by Eltek, which is normally not provided to end
users. However, Eltek is providing UiA-Grimstad with a licence until the end of 2015 [19].
Appendix H shows the full price quotation from Eltek for the Theia HE-t inverter to be 6840
NOK plus taxes and transportation costs.
The DLX 3.8 is produced by Danfoss in Denmark [64]. The DLX 3.8 has a maximum efficiency
of 97.3 % and a casing manufactured after the IP65 protection level. It uses an HF transformer,
providing galvanic isolation and giving it a total weight of 23 kg. Included in the inverter is
reverse PV polarity protection, ground fault monitoring, integrated DC switch and excessive
PV power protection. The DLX is a robust inverter that can operate at an ambient temperature
range between
colour display [65]. A quotation was given from a Danish company called Soltek for 9800 DKK,
equivalent to approximately 10900 NOK (25/2-14), plus taxes and transport [66].
The Fronius IG plus 35 is also an inverter supplied by Getek. The design of this inverter is
an enhancement of other successful inverters with improvements to provide higher efficiency
and earnings [67]. A combination of multiple power stages provides maximum power output
on cloudy days. It operates with a HF transformer with transformer switchover, resulting in a
constant efficiency over a wide range of the input voltage [67]. The Fronius IG plus 35 operates
between
20 C and 55 C and has a weight of 23.8 kg. It has protection degree IP54, meaning
that it is dust proof and tested against splashing water [68]. The data logging capability of
Fronius IG plus 35 is not specified, but an extra data logger/manager can be installed to
overview how the system operates at all times [69]. A price quotation was given by Getek for
the Fronius IG plus 35, they oered 10 000 NOK plus taxes and transportation [70].
Appendix C shows the inverters found with the appropriate power rating and the most important parameters for each inverter. For a PV system to be approved and connected to the grid,
the local utility company needs to approve the system, inverter and the state of the power output. Agder Energi writes in their regulations [71] that they highly recommend having galvanic
isolation in the PV inverter and specifies that it is not allowed to leak any DC current into the
grid.
The inverter with no transformer, PVS300 TL 4000 from ABB and SMA 3800 because of a
lower maximum DC system voltage and a datasheet that does not give satisfactory information,
is ruled out. The three inverters remaining are Theia HE-t 3.8 from Eltek, DLX3.8 from
Danfoss and Fronius IG plus 35. The inverters from Eltek and Danfoss have the same technical
40
properties, and power rating. In addition they have the same protection degree (IP 65), number
of string inputs and efficiency. They both have an integrated DC switch, reverse polarity
protection and ground fault monitoring. The dierence occurs when the price is considered. As
explained earlier, the model from Eltek has a price of NOK 6840 before taxes and delivery cost,
this is a quotation given especially for UiA Grimstad. Soltek supplies the Danfoss DLX 3.8 for
10900 NOK before taxes and delivery cost. One can assume that the delivery cost will be higher
when transporting 21 kg from Denmark than from Drammen, in addition customs and taxes
has to be paid when transporting it over the border. The Fronius IG plus 35 has protection
degree IP54, meaning that it is less dust- and waterproof than the other two. It also does not
have the same protective equipment, it is only equipped with a DC switch, overload behaviour
and DC insulation measurement. The data are also only available if an external data manager
is bought.
When considering these factors and the fact that Eltek is a local company relatively close to
Grimstad, the Theia HE-t 3.8 is considered to be a good and reasonable inverter for this small
PV system.
5.2
The power quality requirements are written in the Norwegian law, Forskrift om Leveringskvalitet i kraftsystemet (FoL), and give the Norwegian water- and energy directorate authority
to impose people or companies that fully or partly owns, operates or use electrical installations
or electrical utilities which are connected to the Norwegian utility grid and the person that
the energy law defines as system operator [72], to do the necessary measures to reduce the
consequences of short- and long-term power outage [72]. FoL and Agder Energi use dierent
requirements to ensure the power quality, since Agder Energi uses dierent values to ensure
that the requirements in FoL are never broken, their requirements are considered in this thesis.
These requirements for power quality as stated from Agder Energi in Appendix J and [71] are
also applicable for plus customers and eects the following terms:
Slow variations in the voltage RMS value
Step voltage Changes
Flicker and Flicker Severity
Harmonic Distortion
5.2.1
41
Slow variations in voltage are defined as changes in the stationary voltage RMS value over a
period of time. To avoid significant voltage variations at the end consumers, it is not allowed to
produce a voltage that during an average time of 1 minute exceeds 10 % of the nominal RMS
voltage [72]. The nominal RMS voltage can therefore only vary between 207 V and 253 V for a
230 V grid, as illustrated in Table 5.1, but cannot exceed these limits for more than 1 minute.
This is a limit programmed in the inverter before it is connected to the grid and can be changed
dependent of the requirement that apply.
Table 5.1: Allowed voltage variations in the grid over 1 minute [72].
Voltage level
230 V
400 V
5.2.2
The step voltage is defined as rapid changes in the voltage RMS value and is caused by errors
and short circuits in the utility grid or big loads at the end user. Step voltage is variations or
steps between 207 V and 253 V happening faster than 0.5 % of the agreed voltage level per
second [72]. E.g. for 230 V grids, the voltage cannot change more than 1.15 V per second. If
this happens, FoL has put limitations shown in Table 5.2 for how often it can happen in grids
with a nominal voltage between 230 V and 35 kV.
Table 5.2: Number of times the step voltage, expressed as Ustationary and
deviate the limits for one 24-hour time period [72].
Step voltage
Ustationary 3 %
Umaximal 5 %
Umaximal , can
The step voltages in Table 5.2 are limits for when a step is counted, the value connected to this
limit is how many times this can happen during a period of 24 hours. FoL is expressing step
voltage as stationary and maximal change in voltage as given:
% Ustationary =
Ustationary
100 %
Uagreed
(5.1)
42
% Umaximal =
Umaximal
100 %
Uagreed
(5.2)
5.2.3
Flicker is defined as the visible variation where the luminance (candela/m2 ) or the spectral
distribution varies with time [72]. Flicker is the visible variation of light caused by RMS
voltage variations with a frequency between 5 and 20 Hz. The worst frequency for the human
eye is statistically 8.4 Hz, this leads to discomfort for people because of intensity variations in
light [73]. Flicker severity is the intensity of the flicker discomfort and is defined by the UIE-IEC
flicker measuring method. There are two indicators for determining dierent levels of flicker,
according to FoL [72].
i Short term flicker indicator (Pst ), where the flicker severity is assessed after a time period of
10 minutes.
ii Long term flicker indicator (Plt ), where the flicker severity is calculated over a time period of
two hours using 12 Pst numbers calculated from Equation (5.3).
v
u 12
uX P 3
3
st i
Plt = t
12
(5.3)
i=1
where Pst
Energis technical requirements, as illustrated in Table 5.3 from Appendix J, should be followed
when choosing an inverter. The inverter makers should have set design limits so the inverter
complies with the requirements set in Table 5.3. Only a technician authorised by Eltek can
change the parameters [75]. A Type Verification Test Sheet of the Theia HE-t made by Bureau
Veritas Consumer Products Services in Germany, states that the inverter complies with standard
43
230 V Unom 35 kV
1.0
0.8
Time interval
95 % of the week
100 % of the week
EN 61000-3-3 [75]. During the testing at 100 % power a Pst and Plt of 0.43 was measured, which
is well below the limits required by Agder Energi in Table 5.3.
5.2.4
Harmonic Distortion
If an appliance does not have a linear relationship between current and voltage, then this
appliance will be a source of harmonic distortion. The PWM control in an inverter is also
a source of harmonic distortion if no extra capacitors have been added to reduce this. If
powered by a sinusoidal voltage, like Figure 5.1 shows, and the current is not sinusoidal then
harmonic distorted currents will occur. Examples of appliances that create harmonic distortion
are power electronic transformers, electric arc furnaces and welding equipment [76]. If these
harmonic distortion currents are mixed with large non-linear impedances in the utility grid,
then harmonic distorted voltages occur [77]. The inverter will for PV systems be one source of
nonlinear current and voltage characteristic, and could cause harmonic distortion in the utility
grid. An example of a sinus distortion at third harmonic is shown in Figure 5.1. Harmonic
44
FoL does not set any other requirements to inter-harmonic than providing the Norwegian waterand energy directorate the authority to decide inter-harmonic limits at the connection point.
A specific value for this has not been determined and inter-harmonic voltage is therefore not
included in this report. There are usually several harmonic components in a grid. A way to put
these together is to use the Total Harmonic Distortion (THD) as defined in Equation 5.4.
%T HDh =
P40
2
Uh2
U1
100 %
(5.4)
U1 is given as the basic harmonic component of the voltage, and the individual harmonic distortion, Uh , for every single multiple of the basic harmonic frequency is defined in Equation 5.5
[72].
% Uh =
Uh
100 %
U1
(5.5)
FoL state that it is the utility companys responsibility to ensure the total harmonic distortion
(THD) does not exceed 8 % and 5 %, measured as an average over 10 minutes and one week
respectively, at the connection point where the nominal voltage is between 230 V and 35 kV [72].
To comply with this, Agder Energi operates with stricter requirements stating that the THD
should not exceed 6 % and 4.5 % as an average over 10 minutes and one week [71], respectively.
It is also the utility companys responsibility to ensure that the individual harmonic distortion
(Uh ) does not exceed the values in Table 5.4, measured as an average over 10 minutes, at the
connection point where the nominal voltage is between 230 V and 35 kV [72]. In addition,
Table 5.4: Odd- and even-harmonic distortion limits as stated by Agder Energi in Appendix
J.
Odd Harmonics
Not multiple by 3
Order h
5
7
11
13
17
19
19
23
25
> 25
Even Harmonics
Uh
5.4
4.5
3.2
2.7
1.8
1.4
1.4
1.4
1.4
0.9
%
%
%
%
%
%
%
%
%
%
Multiple by 3
Order h
Uh Order h
3
4.5 % 2
9
1.4 % 4
15
0.5 % 6
21
0.5% > 6
> 21
0.5 %
Uh
1.8%
0.9%
0.5%
0.3%
Agder Energi has a limit for harmonic currents stating that the PV installation must meet the
limit in Table 5.5 for relative harmonic currents given in IEC 61000-3-6 [71].
45
Table 5.5: Limits for relative harmonic currents as stated by Agder Energi in Appendix J.
Over-harmonic order h
5
7
11
13
qP
i2h
Over-harmonic current ih =
5-6 %
3-4 %
1.5-3 %
1-2.5 %
Ih i
[%]
Ii
6-8 %
It is Agder Energis right to impose stricter requirements than stated in FoL. This is done so
the requirements in FoL can be fulfilled in grids with several local production units. As shown
in Table 5.4 and 5.5, requirements are given for harmonic distortion because hosting several
sources of harmonics could cause requirements in FoL to be exceeded.
IEC 61000-3-6, which is a requirement from Agder Energi that the inverter must fulfil, is an assessment of emission limits for the connection of distorting installations to medium voltage,
high voltage and extra high voltage power systems [79]. Inverters from Eltek are only
connected to low voltage grids (< 1 kV ) and are therefore not covered by this standard. The
equivalent standards IEC 61000-3-2, limits for harmonic current emissions (equipment input
current 16 A per phase) [80] and IEC 61000-3-12, Limits for harmonic currents produced
by equipment connected to public low-voltage systems with input current >16 A and 75
A per phase [81] cover the requirements for harmonic distortion as listed in EN 50438 [82].
The type test verification sheet [75] also provides data for harmonic distortion, which is well
within the limits set by Agder Energi. The Total Harmonic Distortion (THD) of Theia HE-t is
measured to be approximately 3.65 % [82], which is within the requirement set by Agder Energi.
5.2.5
Discussion
The testing results of the Theia HE-t 3.8 in [75] provides a value for the requirements given
in Appendix J from Agder Energi. The inverter in [75] is tested for UK grid, but some values
(like harmonics) are standard for the inverter and therefore common for UK and Norwegian
grids. It states that, for the power quality requirements in Appendix J, the inverter from Eltek
is sufficient and complies with the required standards for the Norwegian grid. However, this
is not surprising since Eltek produce inverters in Norway that complies with European and
international standards, it is assumed that they also build inverters for use in Norwegian grids.
The main reason why requirements are given for power quality from solar inverters is to ensure
the total power quality of the Norwegian utility grid. Deviations from this would result in an
46
unstable grid, which could destroy the electrical appliances it is meant to supply. Theia HE-t is
tested towards these standards, which was made to prevent significant deviations, and is found
adequate to be used in both Norwegian and German (and other) utility grids.
5.3
5.3.1
It is called step voltage if the voltage varies between 207 V and 253 V with the limitations
explained in Chapter 5.2.1. If the voltage exceeds the 10 % nominal voltage limit, it is
considered as a voltage dip or an undervoltage less than 207 V and swell or an overvoltage
when the voltage is above 253 V [83] [84]. A voltage dip is explained as a rapid reduction of the
nominal RMS voltage to below 90 % and higher than 5 % at a duration between 10 ms and 1
minute. Voltages below 5 % of the nominal voltage are considered as an electrical outage [72].
A voltage dip is mainly caused by short circuits, reconnection when errors occur and when large
loads are added to the grid [77]. A voltage swell in an increase of the nominal RMS voltage by
more than 110 % of the agreed voltage level within a duration between 10 ms and 1 minute [72].
A voltage swell does not occur often and is mainly caused by rapid disconnection of large loads
or wrong transformer stepping [84]. Figure 5.2 illustrates voltage dip a) and voltage swell b).
The limits of 253 V and 207 V are illustrated with a red and blue line. If a voltage dip or swell
occurs, Agder Energi states that the PV unit must be automatically disconnected within the
time frame given in Table 5.6. The table show voltage dip and swell in percentage related to
the nominal voltage (UN ). It illustrates that if the voltage deviates within 10 % of the nominal
voltage, the inverter can use 60 seconds to disconnect from the grid. If the voltage deviation
exceeds 15 %, it must be disconnected within 0.2 seconds [71].
47
Figure 5.2: (a) A voltage dip caused by a short circuit in the grid (b) A voltage swell caused
by a disconnection of a large load [83], [84].
Table 5.6: Allowed disconnection delay in case of voltage drops or voltage swells as stated in
Appendix J.
Voltage area in % of UN
U>>115 %
U> 110 %
U<90 %
U<< 85 %
5.3.2
Abnormal Frequency
The Norwegian utility grid is very stable and coordinated, the frequency is therefore not the
parameter that tends to cause problems. Problems could occur in smaller isolated grids not
operated by the utility company (micro grids), or grids that run on generators in general [77].
It is the system operators responsibility to ensure that the system frequency at all times is
within the limitations 50 Hz 2 % (49-51 Hz) [72]. If the frequency would go beyond this limit
measures like disconnecting generators, frequency control of active production and start of gas
generators will be made to control and keep the frequency within these limits [74]. Agder Energi
states that if the frequency deviates from the allowed limit, the inverter must be automatically
disconnected within the time given in Table 5.7.
48
Table 5.7: The required disconnection time at an abnormal frequency as stated in Appendix
J.
Frequency [Hz]
f > 50.2
f < 47
5.3.3
Reconnection
The requirements above are given in case a fault occurs and the inverter needs to be disconnected
to protect the power quality of the utility grid. The inverter can then be reconnected after the
fault is solved; depending on how long the fault lasted. The time frame is given in Table 5.8.
Table 5.8: Delay for reconnection of the inverter after a grid failure has occurred, as stated
in Appendix J.
Fault duration
<3s
>3s
5.3.4
Discussion
The type verification test sheet [75] tested the inverter for the United Kingdom grid code, which
has dierent tripping times than required from Agder Energi. It is also observed that the overand under voltage and frequency are dierent in the UK grid. At first glance, it might look
like the inverter does not fulfil the protective requirement from Agder Energi. However, since
the tripping times are tested with UK settings it may be assumed that Eltek, a Norwegian
company, also has a setting installed for Norwegian grids. The dierent settings possible to
choose are not stated either in the user [53]- or installation manual. If the inverter does not
have Norwegian grid settings, then the procedure from page 75 in the user manual is used: The
settings can be changed within 5 hours after Start up and feed-in to the grid using the Owner
password. Thereafter it is only accessible by using the Installer password, which will only be
available to the installer by contacting Eltek [53]. This gives the impression that the voltage
and frequency, as well as its tripping times and reconnection time on page 76 in the user manual,
can be changed manually within the first five hours after start up and feed-in. From this, it is
assumed that the inverter settings can be manually configured to comply with the protective
requirements given by Agder Energi during the installation and set up of the inverter.
5.4
49
Designing a PV system
Norwegian standards are not completely developed for PV systems since this is still a relatively
new phenomenon in Norway. The German [4] and Australian [3] design and installation manuals
are therefore used where the Norwegian standards lack guidelines. For the technical standards
regarding AC cable CCC, cross section area etc., the Norwegian standard NEK 400:2010 is
used. The inverter chosen for this system is Theia HE-t 3.8 from Eltek with the following
specifications listed in Table 5.9 from Appendix A. One important factor is the maximum DC
system voltage of 600 V, this is the absolute maximum value and must not be exceeded anytime
during operation. The maximum PV power is the recommended maximum PV power at STC,
a value set by Eltek. If more than 5000 Wp is produced by the PV array, the inverter will
only deliver the maximum power and not process the excess power [53]. The PV system will
however, be so small that this will not be an issue. The recommended value is therefore used
as the maximum input power for this PV plant.
Table 5.9: Technical specifications for Theia HE-t 3.8.
Specifications
230 V - 480 V
600 V
18 A
5000 Wp
4000 W
The PV module chosen to constitute the array is the NAPS (NESTE) NP100G12. These are
aged PV modules from the former National Renewable Energy Centre (Energiparken) where
the power ratings have been degraded after spending 10-12 years in the field. The approximate de-rated values in Table 5.10 were measured by our supervisor, Dr. Georgi H. Yordanov,
from outdoor measurements on a single module and re-used in this thesis. The corresponding
temperature coefficients are listed in Table 5.11.
5.4.1
Array Sizing
When designing a grid-connected PV system the properties of both the chosen PV module type
and the chosen inverter type have to be taken into account. In order to produce at optimal
power output the array has to be matched to the inverter. The following steps are used to size
the array [3].
50
Table 5.10: Original and de-rated specifications for the degraded NAPS (NESTE) NP100G12
modules.
NP100G12
Max power
Typical IM P
Typical VM P
Typical ISC
Typical VOC
Operating temperature
Maximum system voltage
Maximum load [Pa ]
Original specifications
100 W 10 %
6.0 A
16.7 V
6.7 A
21.6 V
40 C to +85 C
600 V
N/A
De-rated specifications
92 W 3 %
5.48 A 3 %
16.8 V
6.05 A
21.7 V
N/A
Table 5.11: Temperature coefficients for the degraded NAPS (NESTE) NP100G12 modules.
Temperature coefficients
VOC
VM P
ISC
PM AX
Rating
0.075 V / C
0.0757 V / C
+0.004 A/ C
0.44 W/ C
51
ST C
mod
= VM P
ST C
(T
T0 )]
(5.6)
is the temperature
coefficient of VM P , T is the module temperature at the maximum ambient air temperature and
T0 is the temperature at STC [3].
VM P
mod
= 16.8 V
[0.0757 A/ C (76.4 C
25 C)] = 12.91 V
(5.7)
According to GSES (on page 187), a safety margin of 10 % is added to the minimum inverter
voltage due to the following reasons [3]:
1. The inverter may not always be operating at the ideal MPP
2. Allowing for manufacturing tolerances
3. Allowing for lower light conditions (which causes the module voltage to decrease)
4. Allowing a safety margin in case of shading of one or more modules
The module VM P is dependent on temperature, see Equation (5.7), and therefore also irradiance.
When a higher irradiance is inflicted upon a module, the temperature rises and the module VM P
is slightly reduced due to series resistance. At lower temperatures, the module VM P increases.
A safety margin of 10 % is added to count for this variation in array VM P .
The module VM P was calculated by using Equation (5.7), giving 12.91 V . This value must be
multiplied by 0.98 to account for a voltage drop of 2 % (will be explained in Section 5.4.2)
giving: 12.91 V 0.98 = 12.65 V . Second, 10 % is added to the minimum inverter voltage:
230 V 1.1 = 253 V . The minimum MPPT voltage of the inverter is then divided by the module
VM P :
Minimum number of modules =
Vinverter
253 V
=
= 19.99 modules
VMP
12.65 V
(5.8)
The value calculated in Equation (5.8) is always rounded up. Since a safety margin is already
added, the minimum number of modules allowed in one string for the inverter Theia HE-t 3.8
is 20 modules.
The worst-case scenario happens if the maximum temperature gets higher than the temperature used in the calculations, this will cause the voltage to drop below the minimum inverter
voltage. The MPPT is only tracking the VM P of the array within the voltage range specified for
52
the inverter. If the VM P is outside of this range the MPPT algorithm will find a point within its
range that will produce the maximum amount of power. The input current will then be lower
than the PV arrays IM P and the power will be reduced below PM as illustrated in Figure 5.3.
ST C
mod
= VOC
ST C
+ [ (T
T0 )]
(5.9)
is the lowest expected module temperature at the specified ambient temperature. The module
VOC at a given module temperature is found by Equation (5.10):
VOC
mod
25) C] = 24.34 V
(5.10)
A safety factor of 5 % is subtracted from the maximum inverter input; this is done to count
for manufacturing tolerances and higher voltages than accounted for [3]. The module VOC is
given at STC conditions, 1000 W/m2 , to use a standardised value for all of the modules. As
mentioned in 4.3, the peak irradiance during winter can reach 800 W/m2 for Grimstad. The
modules VOC would therefore be slightly lower during the winter and slightly higher during
summer with a peak irradiance of 1200 W/m2 . Calculations for VOC are not measured for the
PV modules during winter or summer, the STC value is therefore used along with a 5 % safety
factor to count for higher voltages.
Vinverter = Vmax
input
(5.11)
The maximum number of modules is calculated by dividing this more conservative no-exceed
voltage by module open-circuit voltage:
Number of modules =
Vinverter
570 V
=
= 23.418 modules
VOC mod
24.34 V
(5.12)
53
This number is always rounded down to avoid producing over the maximum input voltage. The
maximum number of modules allowed in the string is therefore 23 modules.
ST C
mod
= ISC
ST C
+ [ (T
T0 )]
(5.13)
mod
25 C)] = 6.25 A
(5.14)
This is the module short circuit current at STC irradiance conditions and a high module temperature. It is possible to exceed the extra-terrestrial irradiance in the southern parts of Norway.
A research paper published by UiA [30] states that the maximum expected overirradiance in
short bursts for Grimstad is 1600 W/m2 .
ISC
mod
25 C)] = 9.89 A
(5.15)
A higher overirradiance in Grimstad would provide a higher module short circuit current of
9.89 A. If this system was to be designed for operation at the Equator, a short term irradiance
of 2000 W/m2 could be possible, which would give a module short circuit current of 13.5 A.
This would increase the array current and reduce the number of strings possible to attach to the
inverter as well as aecting the cross section of the chosen cables. The German design manual
[4] states the maximal values for the irradiance in central Europe to be 1500 W/m2 , which is
a pessimistic number when the overirradiance in Grimstad could reach 1600 W/m2 . It is well
known that the irradiance increases at lower latitudes. The German manual lacks to investigate
how overirradiance at lower latitudes both in Europe and by the Equator would increase the
module short circuit current at short intervals of time, like in Figure 4.1. The consequence of
not counting for these overirradiance peaks would be the loss of power when the current exceeds
54
the MPPT-range.
Number of strings =
Imax DC input
18 A
=
= 2.88
ISC mod
6.25 A
(5.16)
The ISCmod of 6.25 A is used further because this is instructed in the manuals, and because
the PV system is relatively small. The maximum number of strings is calculated by Equation
5.16 and is always rounded down to avoid over current in the inverter, the maximum number
of strings allowed in the array is two strings. This is not consistent to the number of string
input of the inverter given in Appendix A, which is three strings. It seems like the inverters DC
current rating is optimised for two strings of PV modules with higher ISC up to 8.5 A (equals
2.1 strings) and 3 strings of PV modules with a lower ISC down to 5.8 A (equals 3.1 strings).
Other PV modules with an even higher ISC could also be installed if only one string is used.
(5.17)
This is rounded down and means that 54 modules can be connected to the inverter. Current
and voltage calculations show that this is not possible, a maximum value of 23 modules in a
string and maximum 2 strings give a maximum number of 46 modules in the array. The power
calculations are necessary to ensure that a PV system is not oversized. If that happens, the
inverter will not be able to process the power shown in Figure 5.3 and will instead limit the
produced array power by creating a gap between the arrays MPP and the operational range of
the MPPT [3].
55
ISC
Pmp
Current
Vmp, Imp
Power
Voltage [V]
VOC
Figure 5.3: Illustrates how some of the arrays power cannot be processed because it is outside
the MPPT range of the inverter [3].
Discussion
Calculations done above are guidelines on how the array should be sized when following the
design manuals. It is therefore chosen to follow this and not count for overirradiance, since it is
not specified in the manuals, even though it is known that overirradance of 1.6 suns can occur.
The system will be oversized, so further results when the inverter starts commissioning will
explain how overirradiance aects the system. Because of limited space, the mounting system
on the roof of UiA Grimstad can only fit about 32 modules. With the calculated limitations
of a string containing between 20 and 23 modules and a maximal number of two strings, it
can be concluded that there will be no room for two strings with 20 modules each. The best
option would be to install one string with 23 modules in series to make the most out of one
string. This will give a rated system power of 2.1 kWp before system losses in Section (5.5)
are accounted for. The 23 modules will be arranged in the configuration given in Figure 5.4 to
minimize losses due to shading from the fence and expose as much of the array as possible to
the sun. This figure illustrates the order of how the modules are mounted, the polarity, how
they are connected to each other in series and will be used later in the thesis. The figure also
give an equation for calculating the length of the shadow from the fence in the plane of the PV
array at sunrise and sunset, as height of fence/sin(40).
5.4.2
Cable Sizing
Wiring is an important part of the PV system design, both for safety- and efficiency reasons. It
is important that the conductor and the electrical insulation is correctly sized since undersized
AvanCIS
-6+
-5+
-7+
-9+
- 10 +
- 11 +
- 12 +
- 13 +
-8+
Gllivare
TC
+ 17 -
+ 16 -
+ 15 -
+ 14 -
- 20 +
+4-
+3-
+1-
- 18 +
- 19 +
+2-
+ 23 -
+ 22 -
+ 21 -
1.17 m
Fence/sin(40)
a-Si
56
11.45 m
Figure 5.4: PV array configuration as it was installed on the roof, showing the order and
polarity of the PV modules as well as the four elongated string cables.
cables could cause a fire hazard. If cables are correctly sized, the voltage drop will be minimal
and current in cables will never be greater than the cable safe current handling capability. There
are dierent currents and voltages in dierent parts of the PV system. PV system cables are
therefore divided into string-, array- and AC- cables [3].
The voltage drop in these cables can not be so large that the electrical appliances do not get the
power it is rated for at normal operation [85]. AS/NZS5033:2005 states the DC voltage drop
to be less than 4 % to allow for an AC voltage drop of 1 % [3], while the Norwegian standard
NEK 400-5-52:2010 paragraph 525 states that the total voltage drop should be between 5 % to
8 %. The Australian standard, which was made for PV systems, is therefore applied. The 4 %
DC voltage drop is divided into 2 % for the string cables and 2 % for the array cables.
Cables with steel or aluminium conductor are sometimes used in electrical plants [19], but
copper has the best conductivity. The cable used in this PV system is a 6 mm2 solar cable with
copper conductor, a DC rating of 1.8 kV and a rated current of 32 A [86]. UiA has this cable in
storage and we will use this when extra cable length is needed instead of buying a new cable.
57
(5.18)
7.81 A
(5.19)
In this section it was decided that only one string is to be used, the array current given as
ISC Array equals one modules rated current, which equals the string current. The peak continuous irradiance measured in Grimstad is about 1200 W/m2 , which can last for tens of minutes
[19]. It is therefore assumed that a CCC of 1.25 times ISC can be used [4]. The array CCC
found with Equation (5.18) is therefore the same as for the string cables. The current carrying
capacity for the AC cable is found in table 52B-2 to 52B-13 in the Norwegian standard NEK
400-5-52:2010 for a chosen cable cross section assuming an air temperature of 30 C.
PV String Cable
The PV modules are delivered with string cables built in from the factory. These cables are
already dimensioned with the module current to be 2.5 mm2 with a total length of 0.9 m
both ways for the series cables between the modules and are therefore not necessary to resize. However, it is desirable to minimise the loop area of the array as illustrated in Figure
5.5. To reduce lightning induced voltage spikes it is recommended by the Australian standard
AS/NZS5033:2005 to wire the array so that the area of conductive loops is minimal [3]. To
ensure this, string cables should be installed as shown in Figure 5.5(a), where the negative and
positive string cable are at dierent lengths. This would result in a negative string cable length
of 25 metres one way. Copper cables with 6 mm2 in dimension intended to use as array cables,
is used as negative cable and would give an additional voltage drop as calculated in Equation
(4.6).
The wiring loop cable consists of a copper conductor. The resistivity for copper at 20 C is found
to be approximately 0.0168 mm2 /m [87], with a temperature coefficient of approximately
0.43 %/ C in the range between
mm2
0.43 %
= 0.0168
1+
(70
m
100 C
20) C
= 0.020412
mm2
m
(5.20)
58
Figure 5.5: Correct and incorrect wiring of the string cables with regards to minimizing the
wiring loop [3].
The voltage drop over the string cables is then calculated to be:
2
Vdrop
DC
(5.21)
The calculated voltage drop for the series cables already installed between each module of 0.06 V
is equivalent to a voltage loss of 0.01 %. The distance of 0.9 m occurs between each module, the
voltage drop is then multiplied with the number of modules to find the voltage drop of all the
0.9 m cable distance together. Figure 5.4 indicates that four elongated cables must be added
between some modules, these elongated cables are made out of 6 mm2 cables with a distance
of 0.8 m and must be added to the total string cable voltage drop, which is calculated to be:
(0.06 V 23) + (0.02 V 4) = 1.46 V which equals a voltage loss of 0.38 %.
2
Vdrop
DC
25 m 7.81 A 0.020412 mm
m
= 0.66 V
6 mm2
(5.22)
The negative cable length added to minimise the conductive loops results in an additional
voltage drop of 0.66 V, equivalent to a voltage loss of 0.17 %. The total voltage drop for the
string cables is 2.12 V or 0.55 %, when it is assumed no loss in the MC3 string connectors,
which is lower than the recommended value of 2 %.
PV Array Cable
The PV array cable is the connection between the string and the inverter and is often the
longest cable in the system. The voltage drop over the array cable is also assumed to be 2 %,
59
the MPP of the string at STC is expressed as the number of modules times module VM P ,
23 16.8 V = 386.4 V [3]. The array cable will be installed through a duct from the array on
the roof to the inverter located in the PV lab. The total length of the array cable is assumed to
be 50 metres both ways, where 6 metres of the cable is located in open with the sun heating up
the black insulation. It is assumed that the exposed cable temperature can reach 75 C. The
copper resistivity for the exposed 6 metres is then calculated in Equation (5.23).
exposed
mm2
0.43 %
= 0.0168
1+
(75
m
100 C
20) C
= 0.0207732
mm2
m
(5.23)
It is assumed that the temperature for the remaining 44 metre of the array cable, located in a
duct, can get up to 30 C. The resistivity is calculated with Equation (5.24).
array
mm2
0.43 %
= 0.0168
1+
(30
m
100 C
20) C
= 0.0175224
mm2
m
(5.24)
The dierence in resistivity between the exposed cable and the enclosed cable is not very large
and is almost negligible. If exposed were to be used to calculate the minimum cross-section area,
the dierence would also be very small. Even though six metres of the array cable is exposed to
75 C, 30 C is still used as cable temperature because of the small dierence and since 6 mm2
(explained below) is a lot thicker than the calculated minimum cross section area in Equation
(5.16).
2
ADC
array
2 25 m 7.81 A 0.0175224 mm
m
=
= 0.886 mm2
0.02 386.4 V
(5.25)
By using the two dierent temperatures, a minimum cross-section of 1.032 mm2 is reached when
75 C is used and 0.886 mm2 when 30 C is used. The dierence is minimal, especially because
an array cable with a cross section of 6 mm2 is already installed between the mounting rack and
the PV lab, which has been used for other research projects. Since the minimum cable size is
0.886 mm2 , the 6 mm2 array cable is both oversized and sufficient to supply the array current
to the inverter. A new array cable will therefore not be installed. With the array cable cross
section selected, the voltage drop can be calculated with Equation (5.26).
2
Vdrop
DC
2 25 m 7.81 A 0.0175224 mm
m
=
= 1.14 V
6.0 mm2
(5.26)
A voltage drop of 1.10 V over 50 metres is equivalent to a voltage drop of 0.3 % this is also
less than the recommended maximum voltage drop of 2 %.
60
AC Cable
The AC cable length from inverter to electrical switchboard is measured to be 22 metres, the
maximum output current from the inverter is 19.7 ARM S (Appendix A), the AC phase voltage
is 240 VRM S and the power factor (cos ) of 1.0 is used in the case when current and voltage
are in phase [19]. It is assumed a temperature around the AC cable of 30 C since it is located
inside a building. The resistivity of the AC cable is the same as calculated in Equation (5.24)
[3]. The minimum theoretical AC cable cross section is expressed with Equation (5.27):
AAC
LAC
cable
cable
(5.27)
is the length of the AC cable, IAC is the AC current (expressed as the maximum
inverter output current), is the resistivity, Loss is the allowed voltage loss in percentage
written as a decimal and VAC is the AC phase voltage of the grid [3]. The minimum AC cable
cross section is calculated to be:
2
AAC
cable
22 m 19.7 A 0.0175224 mm
m 1.0
=
= 3.164 mm2
0.01 240 V
(5.28)
The actual AC cable cross-section is selected to be 6 mm2 , almost double of the minimum AC
cable cross section calculated in Equation 5.28. The cross section was chosen after method A2
in NEK 400:2010 assuming two conductors in table 52B-2, giving the AC cable a CCC of 32 A
[55]. According to NEK 400:2010, the minimum AC cross section is 4 mm2 [55] which also
meets the calculated cross section in Equation 5.28, but 6 mm2 is used in case a system with a
higher peak continuous current is installed at a later time. The voltage drop in the AC cable
can then be calculated by transposing Equation (5.27) [85].
2
VDrop
AC
22 m 19.7 A 0.0175224 mm
m 1.0
= 1.27 V
2
6 mm
(5.29)
A voltage drop of 1.27 V over 22 metres is equivalent to a voltage loss of 0.53 %. The total
voltage loss between the array and the kWh-meter is 0.55 % + 0.3 % + 0.53 % = 1.38 % < 5 %.
The requirements in AS/NZS5033:2005 and NEK 400:2010 has been fulfilled.
61
Discussion
It is fairly easy to reduce the voltage loss in the system cables. The solution is to choose a
conductor with a larger cross-section than needed, the voltage loss is then reduced and an extra
safety factor is added when a thicker conductor can carry a higher current. In addition to this,
an extra safety factor was also added when the CCC was found by using 1.25 times ISC
ST C .
For this thesis, UiA has already decided the cross section area of the DC and AC side and
bought the cables to be used. An important part of this chapter is then to assure the quality
of the chosen cables, to ensure a sufficient cross section area for the current it is going to carry
and to ensure the total voltage loss is within the acceptable limits. The calculations done in
chapter 5.4.2 states that the cables chosen for this PV system is more than sufficient to carry
both the current produced in the PV string and a higher current for a future upgrade of the
PV system. The AC cable is dimensioned to carry more than the maximum current possible
for the inverter to produce and is also dimensioned with the intention of a future upgrade.
5.4.3
PV System Protection
One of the most important aspects when designing a safe and reliable PV system is the sizing
of fuses and circuit breakers. System protection is needed to minimise the risk of damage to
system components such as cables and modules. A PV system normally requires two types of
protection [3]:
Over-current protection on the DC and AC sides
Disconnect/isolator switches on the DC and AC sides
According to NEK 400-7-712:2010, a disconnect/isolator switch is needed on both sides of the
inverter. This makes it easy and safe to perform maintenance on the inverter. Over-current
protection is used to protect components and cables from short-circuits or overload currents
and is based on the array current (ISC ). However, on the DC side NEK 400-7-712:2010 states
that over-current protection is not needed if the current-capacity of the conductors is larger or
equal to 1.25 ISC .
An isolator switch is included in the inverter for isolating the DC side and using fuses could
fulfil the overcurrent protection. However, it is decided to use a circuit breaker, which does
not need to be replaced when tripped. Both types of protection listed above are included in
62
a circuit breaker. The DC side of the PV system will be protected by a circuit breaker from
manufacturer Eaton and a circuit breaker from ABB will be used on the AC side.
From NEK 400-4-43:2010, two requirements must be fulfilled in order for a circuit breaker to
protect a conductor from over-current: ISC
Array
rated current of the circuit breaker. I2 is the overload-current ensuring that the circuit breaker
trips. On the DC side, 6 mm2 Cu conductors will be used. According to NEK 400:2010, a 6 mm2
conductor can carry up to 32 A at an ambient temperature of 30 C. Installation method A2 in
table 52B-2 (NEK 400-5-52:2010) is used assuming a single-conductor cable installed in a pipe
in a thermic isolated wall. The CCC of the DC cable when using 1.25 x ISC
array
is 7.81 A,
while the chosen oversized cable has a CCC of 32 A. Currents of this magnitude is not assumed
to occur in the PV system, so the circuit breaker chosen from Eaton is 20 A but the CCC of
32 A is used further. The thermal tripping characteristic is from 1.05 to 1.3 In and is found in
the technical data in Appendix D. The thermal tripping characteristic is preferably set to 1.3,
allowing higher overload-current before the circuit breaker trips. I2 = 1.3In = 1.320 A = 26 A.
The specifications are summarised in Table 5.12.
ISC
Array
In CCC ! 6.25 A 20 A 32 A
(5.31)
Circuit breaker
20 A
(5.30)
Conductor size
6 mm2 Cu
63
On the AC side, the maximum possible current load from the inverter is 19.7 A (IB ) and the
conductor size is set to 6 mm2 Cu. Installation method A2 (NEK 400-5-52:2010) is used assuming a multi-conductor cable installed in a pipe within a thermic isolated wall. The conductor
can carry up to 32 A, which is used as CCC in Equation 5.32. The common B characteristic is
used for the circuit breaker (ABB) where I2 = 1.45 In = 1.45 20 A = 29 A. The specifications
on the AC side are summarised in Table 5.13.
IB In CCC ! 19.7 A 20 A 32 A
(5.32)
(5.33)
Maximum AC current
19.7 A
Circuit breaker
20 A
Conductor size
6 mm2 Cu
Discussion
The Eaton circuit breaker, for the DC side, is constructed for PV systems and also for the
varying currents. Characteristics for this circuit breaker are given in Appendix D and provides
both the tripping characteristic curve and the adjustment tool. The Eaton PKZ-SOL20 can
vary its setting value between 14 A and 20 A dependent of the ISC . The lowest value 14 A
is therefore used as the setting value since this is the lowest possible. The circuit breaker is
therefore rated for 2.2 suns x ISC which is considered more than what is needed when the
highest irradiance is 1.6 suns. Since ISC is considered never to exceed 14 A, the circuit breaker
will not trip at overirradiance conditions and the inverter will therefore be able to utilise this
power.
The cables are oversized and over-current protection should not be necessary on the DC side,
due to CCC being larger than 1.25 ISC . However, a circuit breaker will be installed due to
possible future upgrades on the PV system. The modules can then be replaced in the future
with newer more efficient modules without needing to upgrade cables and the circuit breaker.
5.5
64
Theoretical systems may have 100 % efficiency, but this is not the case when installing a real,
physical system. A PV array has several factors that may cause losses in the system. The
factors mentioned underneath provides an estimate of losses on a yearly basis, which will be
deducted from the total power rating [3].
Temperature and irradiance losses
Dirt and soil losses
Manufacturers tolerance
Shading losses
PV module orientation
PV module tilt angle
Voltage drop through System Cabling
Inverter efficiency
Angular reflectance eects
Spectral gain
5.5.1
The module voltage is dependent of the module temperature, as the temperature rises the
voltage drops. From P = I V , the module power also decreases as the voltage drops. Figure
4.3(a) in Section 4 explains how the I-V curve of a PV module is changing with temperature
variations. Wind is also a factor aecting the temperature, no wind data are measured at the
site and will therefore not be taken into account.
To calculate these losses the relative efficiency (REL ) is used, which is a factor dependent of
temperature and irradiance at a location. The relative efficiency decreases with the latitudes as
the temperature increases at southern locations [45]. Since the module energy yield is strongly
dependent of temperature irradiance, which is varying, the relative efficiency is used to cover
losses due to these factors. This is calculated with Equation 5.34 with respect to STC for lower
irradiances, assuming a constant temperature of 25 C, which gives Figure 5.7(a). Figure 5.7(a)
explains how the relative efficiency of an aged multicrystalline silicon module changes with lower
irradiances. It also illustrates the dierence between an aged module and newer modules with
higher curves and better efficiencies. Figure 5.7(b) shows the energy contribution in percentage
65
for each irradiance measured over Grimstad in 2011, indicating that Figure 5.7(a) is calculated
for the most common irradiances at this location. The relative efficiency and module STC
efficiency is calculated with Equations 5.34 and 5.35. REL is calculated over several irradiances
with respect to STC and constant temperature, the results for this calculation over dierent
irradiances is illustrated in Figure 5.7(a).
REL (G, T ) =
(G, T )
PM (G, T ) G0
=
0
G
P0
0 =
P0
A G0
(5.34)
(5.35)
where P0 is the module STC power rating, 0 is the module STC efficiency, G0 is the irradiance
during STC while the maximum power PM and efficiency is related to the in-plane irradiance
G, at any time, and the module temperature T [89].
Figure 5.7: The relative efficiency is calculated with respect to STC, covering the most measured irradiances over Grimstad in 2011 [89].
From Equation 5.34 and Figure 5.7(a), [89] calculated that an average relative efficiency for
aged mc-Si modules of 0.93 can be used.
5.5.2
Dirt or soil can build up on the PV module if it is not regularly cleaned, this could reduce
the output power by preventing some of the sunlight from reaching the cells. A dirt de-rating
factor, fdirt , is added to the total equation depending on the site and operator. From Figures
66
4.9, 5.8(a) and 5.8(b), examples are given from the location of the PV array illustrating that
the array gets quite dirty over time, this can be prevented by regularly cleaning by the operator.
Figure 5.8(b) shows that the array gets covered by pollen in May, which could aect the power
output if not cleaned. Grimstad is a place exposed to a relatively high amount of rain and
it can be assumed that the PV array will be cleaned after a rainy day, but soiling and bird
droppings need to be manually removed. It is assumed that a PV system operator wants as
small losses from dirt as possible and is therefore cleaning the array when needed. The dirt
factor is therefore assumed to be 0.98.
5.5.3
Manufacturers Tolerance
A manufacturers tolerance is usually showed as a percentage after the module power rating
on the module specification sheet. The tolerance given in the specification sheet is 10 %,
though the calculated de-rated tolerance found by [19] in Table 5.10 indicates a tolerance of
3 %. The manufacturers tolerance should be based on the minimum module power which is
92 W -3 %. A tolerance of -3 % (0.97) is used when doing energy calculations for the PV array.
5.5.4
Shading
Shading is a factor with a significant potential of aecting the output power of the module.
This is a hard value to predict and has unique values for every location, dependent of the
surroundings. If an accurate value for the shading loss is wanted, then detailed analyses should
be made on the roof for at least one year to count for summer and winter conditions. This
is however, not done in this thesis. It is observed that in April, the fence in Figure 4.8 will
67
only shade on the bottom right corner of module number 20 in Figure 5.4 for 30 minutes in the
morning when the irradiance is low. More shading factors during the summer occur because
of birds flying past or landing on the modules, as shown in Figure 5.9(b). During winter, low
sun exposes the array to other shading factors. Figure 4.8 shows shading from both the fence
and snow covering the modules. It is also observed that the telecom tower in Figure 5.9(a) can
provide shading early in the morning for two weeks in spring and autumn. This report uses
approximately numbers, assumed over a year for an ideal system with no fence, telecom tower
or trees to shade the array. It is assumed that these losses will be 3 % (0.97).
5.5.5
Solar systems in Germany and Australia are often mounted on rooftops and are therefore not
always mounted with the optimal tilt angle and orientation. For the location in Grimstad,
PVGIS calculated the optimal tilt angle to 38 from the horizontal with orientation towards
south. The support beams for the PV modules are mounted 7 east from south with 39 1
tilt angle from the horizontal [32]. The exact tilt of the array is uncertain, the margin of error is
therefore used to account for the possibility of the array to have a lower or higher tilt angle. By
using PVGIS, the uncertainties due to tilt angle are estimated to be the same as the margin of
error given as one degree. The uncertainty is estimated by looking at the irradiation in PVGIS
for 38 and 40 for then to calculate the dierence. The uncertainty factor and possible loss
due to tilt and orientation is therefore estimated to be 99.5 % (0.995).
5.5.6
68
The importance of reducing voltage drop is explained in Section 4.4. The important part is
keeping the voltage drop within 5 % between the solar array and the AC socket since a voltage
drop of 5 % equals a power drop of 5 %. Because solar cables are less expensive than solar
arrays, cables with a larger cross sectional area can be chosen to reduce the voltage drop [3]. As
explained in section 5.4.2 the total voltage drop of the PV system was calculated to be 1.38 %,
which equals to a factor of 0.986.
5.5.7
Inverter Efficiency
In addition to the power losses on the DC side of the PV system, there are losses due to
electronics and transformer in the inverter in the form of heating and switching. For the inverter
to operate at maximum efficiency it should be installed with sufficient ventilation and not
exposed to direct sunlight [3]. It is not believed that the inverter will operate at the maximum
efficiency at all time. The Euro efficiency of the chosen inverter, at a specific MPP voltage, of
96.7 % (0.967) is therefore used (Appendix A).
5.5.8
Losses due to sunlight reflecting on the module glass plate need to be added. This factor is
dependent on type of module and on the angle of incident, and aects the amount of irradiance
entering the module and has an impact on cell temperature and array power output [90]. This
factor, fref lection , is estimated in PVGIS and given in Appendix N to be 3.1 % (0.969).
5.5.9
Spectral Eects
The spectral eects of the modules vary and are dependent on the meteorological conditions at
the site. These eects are often so small and uncertain that they are often negligible, especially
when the solar spectra data for overcast conditions are limited [90]. This factor, fspec , is an
energy gain to the system and is assumed to be 1 % (1.01) [19].
5.5.10
69
So far, Section 5.5 in this Chapter explained the dierent factors that could aect the power
output of the PV system. It is calculated in Section 5.4.1 that 23 modules are the maximum
number possible to install in one string. With a PV module power rating of 92 Wp , the power
rating is 2.116 kWp for the entire array. When the de-rating factors are determined, the output
power of the PV system with no losses can be calculated with Equation 5.36 for an estimate of
a year:
Eno
losses
= I A 0
(5.36)
With the relative efficiency and Table 4.3 giving the average monthly irradiation in Grimstad
between 2011 and 2013. The monthly average of the three past years, 103.27 kW h/m2 , which
was calculated in Section 4.1.2 is used. In Appendix N, PVGIS assumes a monthly irradiation
of 86.6 kW h/m2 which is lower than the irradiation used. PVGIS uses PVGIS-3 database in
Grimstad, which is based on on ground measurements and as mentioned in Section 4.1.2 is
known to estimate lower irradiance than the new PVGIS-CMSAF based on satellite data. It
is therefore assumed that the irradiation measurements for the past three years in Grimstad
can give an idea of the past and future irradiation at this location. The yearly energy when
no losses are added is calculated with Equations (5.36) and (5.35), giving 0 , and resulting in
Equation (5.37):
103.27
Eno
losses
m2
kW h
months
12
2.116 kW
months
year
= 2622.23 kW h/year
kW
1 2
m
(5.37)
Equation (5.37) assumes from Table 4.3 that the radiation must continue at the rate of 1 kW/m2
for 103.27 hours per average month to produce 103.27 kW h/m2 [3]. When including other
possible losses in the PV system, as explained in this chapter, a more conservative energy yield
can be calculated.
losses
(5.38)
70
kW h
0.93 0.98 0.97 0.97
year
(5.39)
5.5.11
Performance Ratio
The performance ratio (PR) is calculated to estimate the quality of the PV system and to see
how much the de-rating factors aect the total performance of the system. The theoretical
maximum energy yield is based upon a system with no losses, which is calculated in Equation
5.37.
PR =
(5.40)
2087.83 kW h/year
100 % = 80 %
2622.2 kW h/year
(5.41)
PR =
The performance ratio for the PV system is calculated to be 80 % with the system loss assumptions done above. Analyses done by [91] reviewed the performance ratio in 11 countries
between 1998 and 2002, where the Performance Ratio data bank had recorded PR between 40 %
and 85 %, giving an average mean annual PR of 74 %. From [91] it is concluded that the PR
spread of PV systems installed between 1998 and 2002 had significantly increased compared
to 1991-1998. The PR values from this report are old, but it still expects that the PR of a
well designed PV system will increase in the future. This increase could be a result of better
planning, technology and inverters. The average PR value of 74 % was recorded at the same
time as the PV modules was made; this value is therefore compliant to the age of the installed
71
modules. A PR of 80 % is considered a good value for this system considering the age of the PV
modules which aects the relative efficiency. This value is variable, and will increase for instance
after rainy days, when the PV modules are clean, during the winter when the PV modules are
not covered with snow and when the module operating temperature is lower than the constant
STC temperature.
5.6
PV modules
Specifications of solar cells and modules are determined by using the following standard test
conditions (STC): cell temperature = 25 C; irradiance = 1000
W
;
m2
lack of a solar simulator it was not possible to perform measurements on the PV modules at a
controlled environment with STC conditions. All of the PV modules were therefore tested in
outdoor conditions. To find the best 23 modules out of 48 modules, each module was tested
with an ESL-solar 500DS electronic load connected to the PV module with a positive (red)
and negative (black) cable as shown in Figure 5.11(a). The I-V curve measurements were done
on clear sky days, with an irradiance varying between 595 W/m2 and 1147 W/m2 (noon) and
an ambient temperature varying between 11.2 C and 32.2 C. The conditions were relatively
stable for each module and several of the modules were close to STC temperature. Figure 5.10
shows the temperature variations for module 30. The measurements lasted for approximately
30 minutes for each module.
28
27.5
T [ C ]
27
26.5
26
25.5
25
24.5
0
10
20
30
IV Curve Nr. [ ]
40
50
60
Figure 5.10: The temperature for module 30 changes throughout the testing time and is
illustrated over the amount of measurements done for this specific module.
5.6.1
Testing Equipment
When the Electronic Load ESL-Solar 500DS is connected to the PV module, it uses a load
that is working in MPP tracking mode. In order to stay within the MPP the load works in
(a) The
tracker.
ELS
solar
72
I-V
Figure 5.11: Set-up used when testing the available PV modules with ESL-Solar 500DS.
a continuous mode [53]. To retrieve a good power curve reading, the current and voltage is
measured simultaneously in an I-V curve. The data is transferred from the Electronic load to
a computer via a USB-port where the ESolar software scans I-V curves and plots the MPP
I-V and P-V curve three times every minute. The Electronic Load uses an external sensor for
measuring ambient temperature and irradiance, as shown in Figure 5.11(c), to compare with
the measurements collected from the PV module. As Figure 5.11(b) shows, the solar string
cable is connected to the Electronic Load by using two crocodile clamps, this is one source of
error since the increased series resistance from the clamps could result in less accurate data.
5.6.2
Conversion to STC
The modules were rated according to current, voltage and power. Further on, the 23 modules
with the highest ratings were then selected and used in the PV system. As seen in Figure 5.12
a high number of I-V curves were measured, approximately 50 curves for each module. The
6
Current [ A ]
0
0
10
12
14
16
18
20
22
Voltage [ V ]
solar irradiance and the ambient temperature varied during the measurement period and in
order to compare the PV modules it was necessary to convert the current and the voltage to
73
STC conditions. The short-circuit current ISC [92] and the maximum power current IM P was
converted into STC by using Equations (5.42) and (5.43).
ISC
IM P
ST C
ST C
ISC
G0
GS
1 + ISC (T
IM P
(5.42)
T0 )
G0
GS
1 + ISC (T
(5.43)
T0 )
ISC , IM P , GS and T are the measured values; G0 is the STC value; ISC is the relative
temperature coefficient; T0 is the STC value. The relative temperature coefficients of the PV
modules were measured in outdoors conditions on one module, by determining the rate of
change of current and voltage with temperature. Typically these measurements are done indoor
in temperature-controlled environments and by the use of a solar simulator illuminating the
PV module [92]. However, such a test fixture is not available at UiA. The relative temperature
coefficients are given in Table 5.14.
The open-circuit voltage VOC and the maximum power voltage VM P was converted to STC by
using Equations (5.44) [92] and (5.45) [32].
VOC
VM P
ST C
ST C
VOC
=
1 + V OC (T
= VOC + n Ns V0 ln (1
V OC
T0 )
+ ln
r ) + n Ns V0 ln
G0
GS
G0
GS
(5.44)
r ISC RS
G0
GS
(5.45)
dependency coefficient determined by Equation (5.46); RS is the series resistance and was
retrieved from the I-V curves by using linear regression.
Ns n k (T + 273.15 C)
q
(5.46)
In Equation (5.46) Ns is the number of solar cells; n is the ideality factor; k is the boltzmann
constant; q is the elementary charge. In Equation (5.45) r is the ratio
IM P
ISC
ST C
ST C
A large amount of data was collected during the measurements and a MATLABr script developed by our supervisor was used to perform the STC conversion. Due to asymmetrically
distributed data, median values were used. In order to have as high power ratings as possible,
modules were compared at the maximum point. When series-connecting modules in a string,
the module with the lowest current IM P is operating as the bottleneck of the system. The
74
Table 5.14: Relative PV modules temperature coefficient, number of solar cells in series and
ideality factor.
Parameters
ISC
V OC
NS
n
Value
46.61 10 4 1C
3.46 10 3 1C
36
1.4
Comments
Relative temperature coefficient of ISC
Relative temperature coefficient of VOC
Number of solar cells
Ideality factor
Table 5.15: 28 of 48 highest rated PV modules. Red and orange indicates modules with
low current, voltage and power values. Yellow highlights the current, which will act as the
bottleneck of the PV system. PV modules enclosed by the horizontal lines replace the modules
with values highlighted in red and orange.
Module
1
18
19
3
37
22
29
35
30
38
23
14
40
20
28
36
42
21
5
43
15
34
31
32
27
24
16
41
IM P [A]
5.755
5.706
5.690
5.672
5.646
5.610
5.606
5.575
5.570
5.561
5.555
5.550
5.541
5.534
5.533
5.530
5.524
5.514
5.510
5.496
5.492
5.487
5.486
5.485
5.482
5.481
5.480
5.476
VM P [V ]
15.689
16.567
16.622
15.278
16.431
16.978
16.786
16.712
16.611
17.211
16.695
16.557
16.799
16.627
16.675
16.802
16.844
16.726
16.181
16.692
16.495
16.480
16.647
16.605
16.791
16.708
16.837
16.770
PM P [W ]
90.302
94.644
94.643
86.619
92.701
95.430
94.312
93.132
92.708
95.595
92.761
91.862
92.991
92.000
92.353
92.831
93.060
92.106
88.982
91.725
90.766
90.537
91.357
91.201
92.338
91.487
92.288
91.966
V [V ]
16.255
17.011
17.071
15.689
16.788
17.260
17.069
16.927
16.822
17.379
16.872
16.720
16.940
16.750
16.810
16.913
16.951
16.806
16.259
16.740
16.532
16.506
16.669
16.626
16.806
16.719
16.848
16.770
P [W ]
89.013
93.149
93.480
85.909
91.929
94.514
93.466
92.689
92.117
95.167
92.392
91.555
92.764
91.722
92.053
92.614
92.824
92.030
89.034
91.665
90.527
90.386
91.280
91.040
92.028
91.554
92.256
91.830
FF
0.674
0.692
0.707
0.642
0.695
0.713
0.708
0.701
0.704
0.716
0.704
0.703
0.711
0.692
0.707
0.708
0.710
0.701
0.685
0.709
0.700
0.695
0.695
0.699
0.706
0.705
0.707
0.716
modules were therefore arranged in descending order of IM P . By examine Table 5.15 it is seen
that sorting by IM P is not enough, due to low power PM P and voltage VM P values on certain
modules. 91 W was set as the minimum power value and five of the first 23 modules did not
fulfil this criterion. Five additional modules were added to the list, and new conditions were
needed to ensure the modules with the highest parameters was chosen. This is done by using
75
20
18
16
Count [ ]
14
12
10
8
6
4
2
0
90.5
91
91.5
92
92.5
93
93.5
94
94.5
95
95.5
PM,STC [ W ]
Figure 5.13: Histogram of STC power values from module 30, median values are used when
ranking the PV modules.
I(V ) ISC
I0 exp
V + IRs
nNs VT
(5.47)
I0 exp
V + IRs
nNs VT
1
(dV + Rs dI)
nNs VT
(5.48)
IM IN ) + nNS VT ln
ISC
ISC
IM IN
IM P
(5.49)
Using the lowest IM P as the string current (highlighted in yellow) will result in a new operating voltage and power for each module. The new operating voltage V was calculated by
using Equation (5.50), where
bottleneck IM IN .
V = VM P +
(5.50)
By using IM IN the power values P was slightly decreased, but the runner-ups have a higher
power output than 91 W. In Table 5.15 the modules with parameters in orange and red text
was replaced by the five runner-ups enclosed between the lower horizontal lines.
The fill factor (FF) in Equation (5.51) is the ratio of the maximum power and the product of
the open-circuit voltage and short-circuit current. We used the F F as an indicator to determine
the quality of the PV modules.
76
FF =
Vmp Imp
Voc Isc
(5.51)
Parasitic series and shunt resistances (RP ) can aect the fill factor. Any decrease over the
lifetime may indicate problems with the PV modules. Bulk resistance of the semiconductor,
resistance from the metallic contacts are some of the contributors to the series resistance (RS ).
A high series resistance results in power loss within the cell and thereby decreases the fill factor
and the power output. Shunt resistance is the resistance across the p-n junction. A low shunt
resistance can cause current to leak and hence reducing the fill factor and decreasing the power
output from the cell [3], [28]. In addition to revealing the quality of a solar cell and how square
an I-V curve is, a high F F is also related with the output voltage VM P . A high FF indicates
a high output voltage. The relation between F F , VM P and V is easily observed in Table 5.15.
The modules with the highest FF values also have the highest voltage output. Values between
0.5 to 0.82 are common for solar cells [86]. There are no noticeable deviations except from
module three with a F F of 0.642. However, this module is replaced by the runner up due to
the low voltage and power values [93].
5.6.3
Discussion
The de-rated PM P and IM P values in Table 5.10 was compared to the converted STC values in
Table 5.15 to observe if they are in the same range as the de-rated specifications. The de-rated
maximum power is 92 W 3 % and have a range from 89.24 W to 94.76 W. All except four
modules (3, 5, 22, 28) are within this range when examining the median power PM P of the 28
listed modules in Table 5.15. The de-rated IM P is 5.48 A 3 % and have a range from 5.32 to
5.64 and five of the modules (1, 3, 18,19, 37) are not within this range. However, we consider
this to be acceptable as the majority 20 of 28 modules are within the de-rated specifications.
The STC conversion methods are based on a shunt-free 1-exponential I-V curve model, given
in Equation 5.47. Two assumptions were made in order to derivate the voltage. The shunt
resistance is set to RSH = 1, which is an ideal solar cell, but the resistance could be less than
5
10 . The next assumption is the ideality factor, which is assumed to be 1.4, but this is
probably higher at M P P and also varies with the current. A higher ideality factor degrades the
F F and the modules should probably have somewhat lower F F values. The model assumes the
cells to be identical, but the shunt resistance can vary by nearly two orders of magnitude [19],
77
[93]. Although the model is not 100 % accurate the new operating power provided additional
information to determine the 23 highest ranked modules.
5.7
Installation process
When the circuit breakers, cables, PV modules and inverter are chosen, it is time to do the
physical installation of the PV system. An approval is given from Agder Energi to start commissioning before the agreement is finalised, so that the inverter installations required by Agder
Energi could be done while the final details of the feed in agreement is made. Preparations of
the PV system on the DC side can then commence before the electrical installation and commissioning is initiated. The mechanical installation of the PV modules was done by the authors
of this report while all elements of electrical installation was done by certified electrical students
and controlled by certified electricians from Agder El Installasjon.
5.7.1
The dierent steps of the grid application process are explained in Chapter 3.4 and are commented here. The time elapsed from the application form was handed in to the installation
started was longer then expected. The initial application in Appendix B was e-mailed to Agder
Energi on the 24 January, a response was given on the 11 February saying that the application
had been received and being processed. Some more time then elapsed before Kristen Leifsen
on the 12 March was asked to call Agder Energi to get an estimated time of approval. The call
from Leifsen then speeded up the process and a draft of the agreement was sent from Agder
Energi on 17 March. On 9 April, approval was given from Agder Energi to finalise the remaining agreements after the PV system had started commissioning. After this date, a steady flow
of communication has been present between Rolf H
akan Josefsen in Agder Energi and Bjrn
Lindemann for UiA regarding details in the agreement and documentation needed for UiA to
proceed forward after commissioning the inverter. At the time when this thesis is handed in,
the application process has not been finalised. The documents in Appendix L need to be signed
and sent to Agder Energi before a temporary permit is given. This is changed to a permanent
permit after a trial period when Agder Energi decides that the PV system is producing power
with the required quality.
Comments and updates to the five Grid Connection steps in Section 3.4:
78
It is found in Section 5.2 and 5.3 that the chosen inverter complies with the requirements
from Agder Energi in Appendix J.
The required information in the REN application form, given as Appendix B, regarding
contact person and Produksjonsenhet was filled out by Kristen Leifsen and referred
to general information regarding the placement of the PV system. The authors filled
out the information under Tekniske data. The expected numbers for summer- and
winter-production were found using PVGIS prediction for March-October (summer) and
November- February (winter) which was rounded up and down, respectively. The expected
development cost is based on the inverter cost of 6840 kr and expected installation costs of
11 400 kr. Taxes and transportation of the inverter is not included. The expected cost was
divided by the expected yearly power production for a 3.8 kWp system found in PVGIS
of 3040 kWh/year and resulted in a development cost of 6 kr/kW h. The expected yearly
power production found in PVGIS is for a 3.8 kWp system, while the system installed
is 2.1 kWp . The installed system of 2.1 kWp should have been used, but 3.8 kWp was
used instead because this is the maximum possible output from the inverter, giving Agder
Energi a total power they could expect when accounting for future upgrades. PVGIS
values were used because these were the numbers available at that time, and PVGIS
has a larger data set. The expected cost of 18 240 kr is assumed to be too low when
considering that taxes and transportation should be added to this cost. The price would
be greater for regular customers since UiA got a discount on the inverter from Eltek,
and certified electrical students were used to do the electrical connections. A regular
customer would also have to buy PV modules. 12 years old down-payed modules were
used, no extra costs was therefore added in terms of module price. The development cost
of 6 kr/kW h was calculated using the expected yearly production for a 3.8 kWp system
under the assumption that the PV system would only operate for one year, but is too
high to be a competitive profitable system. This cost should rather be based on expected
production for a 2.1 kWp system with a lifetime of some 10 more years (assuming that PV
module lifetime is approximately 25 years depending on the module and inverter lifetime
is approximately 10 years). The new and more competitive development cost, assuming
a production of 1680 kWh/year from Appendix N for a 2.1 kWp system and an increased
cost of 20 000 kr, is then calculated:
Development Cost =
20 000 kr
= 1.19 kr/kW h
1680 kW h/year 10 years
(5.52)
79
A site acceptance test was at a later time rejected by Agder Energi, instead we needed to
have a mandatory Statement of Compliance form of the electrical installation signed by
Agder El Installasjon, as given in Appendix O. Since quality and protection settings in
Appendix J can only be set after the inverter is connected to the grid, Agder Energi gave
UiA permission to perform commissioning before the grid connection agreement had been
finalised.
After the quality and protection settings has been installed in the inverter, the documentation form in Appendix K needed to be completed explaining who did the electrical
installations of the system, and to add a single-line diagram (Appendix G) and the power
and protection settings installed.
As mentioned earlier, a good communication is present between Bjrn Lindemann and
Agder Energi finalising the details of the agreement. When this is done, the final document
called Rammeavtalen, given in Appendix L, has to be signed saying that the customer
has read all of the documents related to the grid connection and that the system is
operating without exceeding the requirements in Appendix J.
5.7.2
Visual Inspection
The 23 chosen modules in Table 5.15 need to pass a visual inspection test before they can be
used to ensure they are free of damages. The modules should be inspected under an illumination
of no less than 1000 lux for the damages listed [19]:
Cracked, bent, misaligned or torn surfaces
Broken and/or cracked cells
Faulty interconnections or joints
Cells touching one another or the frame
Failure or adhesive bonds
Bubbles or delaminations forming a continuous path between a cell and the edge of the
module
Tacky surfaces of plastic materials
Faulty terminations or exposed live electrical parts
Any other conditions which may aect the performance
The PV modules will pass this inspection if they do not have any major visible defects as listed
above. The defects found and its severity level for every module are given in Appendix F. It was
80
observed that some modules had corrosion, as illustrated in Figure 5.14(a), on the cells beneath
the glass which could be a result of moisture behind the glass caused by delamination. Many
modules also had bent frames, which could be caused by rough handling during transportation
or dismounting. These bends could have left open cracks around the module frame allowing
water or moist air to reach the cells. It was observed that most of the cells with corrosion was
corroded in the corner of the module where water typically would gather.
Nearly all of the PV modules had white dots on the cells, and some of them had something
looking like fingerprints on them. This is illustrated in Figure 5.14(b) and is believed to have
happened during manufacturing, but is not aecting the safety of the module or the performance.
In the PV array, 22 of 23 modules had this phenomenon and it is therefore very common, but
it does not seem to aect the array in any way.
Many modules have been degraded to the extent they got brown discoloured lines around the
cells, as illustrated in Figure 5.14(d). This is not believed to be a result of corrosion or hot
spots, but rather a degradation since the brown lines on some cells could cover all four sides
where water typically would not gather. Since the discolouration is located at the edge of the
cells where it does not cover any cell-fingers, it does not significantly aect the power output of
the modules.
A total of 13 modules had the phenomenon illustrated in Figure 5.14(c), it is not known what
this is or how it is caused, but it looks like something had touched the white fingers before the
paint dried and then dragged it over the cell. This may look like a fault produced during the
manufacturing process along with the white dots, maybe from a faulty machine component.
From Table 5.15, showing the chosen 23 modules, the modules with corrosion, white dots,
fingerprints and brown discolouration ends up high enough on the list to be the preferred
modules for the array. This indicates that the faults found on these modules does not aect the
module IM P or VM P enough to exclude them from the best 23, these are rather the modules
with the best IM P and VM P . The faults does not seem to aect the modules, it is therefore
concluded that the faults do not pose any danger to the PV array or the building.
5.7.3
Mechanical Installation
The mechanical installation of the PV modules started on 7 April and lasted for 2 days. The 23
modules were connected to the supporting aluminium beams with nuts, bolts and washers at
81
(d) Brown
stripes
along the edge of the
cells.
four and 6 points where it was possible. The holes in the supporting beam were not necessarily
made for these modules, so new holes was needed to be drilled on the module frame making it
possible to mount the modules and securely fasten them in the configuration given in Figure
5.4. As mentioned in Section 5.7.2, some of the PV modules had bent frames; this does not
aect the overall strength of the module because a new hole was drilled where the structural
integrity of the frame was still intact. Figure 5.15 illustrates the array after the installation of
the PV modules.
5.7.4
Electrical Installation
Two students who are certified electricians did the installation of the inverter and the circuit
breakers on April 22. Figure 5.16 illustrates the inverter and the junction box mounted on a
plywood plate and then connected to two wall main supports. Figure 5.16(b) illustrates the DC
cables marked with red and green insulating tape for + and
the bottom of the junction box from the roof and elongated with MC3 connectors. The DC
cable is then running through an Eaton DC circuit breaker and connected to the DC input of
82
the inverter with two MC4 connectors. The connection diagram: Unearthed system on page
one in Appendix D was used to connect the circuit breaker.
The AC cable from the wall socket is connected to an AC circuit breaker with one phase (brown
cable) and one neutral (blue cable), the phase and neutral cable then exit the AC circuit breaker
and is connected to the AC output of the inverter. Grounding between the two AC cables is
connected to the earth bar (blue) on top of the fuse box indicating that the inverter is grounded
on the AC side.
Figure 5.16: The inverter installed on the wall along with the junction box and wire connections.
5.7.5
83
Before the inverter could be started, two electricians from Agder El Installasjon inspected the
system and the system wiring. The installation process could start after they inspected and
implemented their suggestions and inputs to the system. Their Statement of Compliance,
where Agder El Installasjon declare that the system is complying with the applicable Norwegian
regulations is given in Appendix O. The start-up was done on 24 April, a cloudy day with a
measured VOC of 346 VDC . The list of Checks before start Up listed in Section 4.3.7 in Theia
HE-t User manual [53] was followed before the installation process started. This list ensured
that all cables were properly sized and connected, as well as circuit breaker size, polarity and the
inverter-mounting bracket. When these points had been checked out, Section 5.1.3 in the user
manual was used to start the installation. A minimum voltage of 230 VDC and 184 VAC , and a
minimum power of 7 WDC needed to be present before the inverter would start. Measurements
done before start-up indicated that these conditions were present. The circuit breakers were
turned ON on the AC side first, then the DC side. The inverters start installation program
is a predetermined guide that is automatically initiated when the inverter is supplied with its
minimum power. The parameters listed in Table 5.16 was set and installed.
The inverter power quality and protection values are predetermined according to European
standard EN 50438:2007 and had most of the requirements from Agder Energi already preset.
Two values were set at Not Applicable in the inverter; Slow trip values for Voltage dip and
Swell. The installers password was needed to manually type these values.
5.7.6
Commissioning
Commissioning is described in [3] and [4] to control the cable cross-section, circuit breakers,
polarity and grounding resistance before and during start-up. This is done because Australia
and Germany have a requirement instructing that a commissioning certificate has to be issued
containing certain measurements of VOC , ISC , grounding- and isolation resistance. However, this
is a requirement that is not necessary in Norway, mostly because standards for commissioning
PV systems has not been developed at this time. Any reports, documentation or information
other than what is specified by Agder Energi is therefore not needed.
The inverter start up was done on a cloudy day at 15:00. The Australian manual specifies that
commissioning should be done on sunny clear sky days when larger voltages, currents and power
84
Point
2
3
4
5
6
7
8
9
10
11
13
14
15
16
17
18
Parameter
Language
Time Format
Date
Time
Set Bus ID
Set as Master unit
Grid Configuration
Feed in phase
Plant Apparent Power
Grid Code
Reactive Power Setting
Screen Timeout
Customer Name
Site
Unit Name
Owner password
Set value
English
24h
24 Apr 2014
15:00
99 (default)
No
TN/TT
Unknown
3.8 kW
Norway
VDE 4105 0-13.8
99
UiA
Grimstad
PV1
****
can be recorded to test the PV system. To comply with this, measurements done on 27 April
at clear sky was used to see how the system operated at these conditions. The result was an
operating voltage of 340 V, operating current of 5.4 A and an internal isolation resistance of
65.5 M . The performance ratio of the system with regards to STC is calculated in Section
5.5.11 to be 0.80. Conditions at 27 April were ideal, which resulted in a better performance
ratio at this day when the inverter recorded an output power of 1921 W at 12:00. It is of interest
to study the temperature and irradiation data for that time, as Figure 5.17 illustrates, to see if
the recorded power complies with the expected theoretical output power when no factors like
shading and dirt was present.
The temperature of 32 C (A10156) was recorded as Figure 5.17(a) illustrates a drop in temperature just before 12:00, this could be a result of a wind gust cooling down the modules and
resulting in a peak power at that time. This is also reasoned by Figure 5.17(b), at 0.975 suns,
illustrating that the peak power did not occur at full sun with a temperature of 34 C but rather
at an irradiance of approximately 975W/m2 .
When calculating if the theoretical power from the PV system corresponds with the actual
AC energy being produced, it can be investigated how the inverter performs with regard to the
calculated expected power. To find this, Equation 5.53 is used with the assumption that voltage
40
35
85
459
Titan
New mcSi
TMOD [C]
30
25
20
15
10
5
6
10
12
Time [h]
14
16
18
20
(a) Module temperature for the NESTE modules on 27 April. 459, Titan and New
mc-si are dierent measuring points on the array.
1200
Irradiance [ W/m2 ]
1000
800
600
Clearsky GHI
GHI
mcSi Cell
Pyranometer
Mean GSELF
400
200
0
6
10
12
Time [ h ]
14
16
18
20
Figure 5.17: Temperature and irradiance measurements on 27 April 2014, the highest power
was recorded at 12:00.
drop is 0.986 and the inverter at this time is operating at the euro efficiency (0.967).
PDC = P0 (Tmod
PDC = 2116 W (32
25) C
25) C
G
+ P0
G0
+ 2116 W = 2046.9 W
92 W
1000 W/m2
(5.53)
(5.54)
5.11. Tmod is measured on the back of the modules and G is measured with the PV cells, which
is assumed to include spectral eects and reflection losses. Equation 5.54 takes into account
the loss due to temperature and irradiance. To see the AC output power, the cable loss and
inverter efficiency need to be added, as calculated in Equation 5.55.
PAC = 2046.9 W 0.986 0.967 = 1951.6 W
(5.55)
86
5.7.7
Thermal Imaging
To inspect how the inverter, fusebox, cables and PV array work with regards to thermal heating
during conditions with high currents and voltages, the system was inspected with a thermal
imaging camera for any hot spots or overheating. The photos were taken on a sunny day when
the PV system was operating at its maximal capacity. It was windy at this day, as it often is on
the roof, and it is believed that the PV array is cooled by the wind at these days. Conditions
on the PV array will therefore be dierent on days with no wind, it will be a lot warmer, but
possible hot spots at thermal heating will still occur on these pictures.
Figure 5.18: Thermal images of the inverter taken when operating at a high power [19].
Figure 5.18(a) is taken over the top part of the inverter, where the transformer is located. The
figure illustrates that the temperature is approximately 28.6 C when the array is operating at
a relatively high power. Figure 5.18(b) illustrates the AC connection point of the AC cable in
the inverter. This picture is taken to ensure that no hot spots occur at this point, indicating
that the cable is not properly installed and must be reconnected. The picture indicates some
red colour on the cable close to the connection point, but it is not a distinct red colour and the
temperature is only 26 C. It is therefore considered not to be a severe issue. Figure 5.18(c)
87
gives an overview of the warmest areas on the inverter. It is seen that the warmest spot is the
ventilation on the side reaching up to 34 C. This overview does not reveal any hot spots in
the cables, junction box or inverter, except from the natural hot sources like transformer and
ventilation.
Figure 5.19: Thermal images of the array on a sunny day indicating that bird droppings and
junction boxes are a heat source.
It is believed that the array was to a high extent cooled by the wind. The temperature is
therefore not consistent to what it would be on days with no wind. Figure 5.19(a) is an overview
of module 4,5 and 6 in the array showing that module 5 has a surface temperature of 25.8 C.
The picture does not reveal any large hot spots in these modules other than heat in the junction
boxes, which is also illustrated in Figure 5.19(b), operating at a temperature of 31.7 C.
Reflections from the sun was prominent on this day, which is also illustrated in Figure 5.19(b)
which is seen by the large red and white surface. Figure 5.19(c) is illustrating something
that could happen to all of the PV modules. Bird dropping on the glass will provide shading
on the module cells causing mismatched cells within the module. This figure shows that the
temperature at the bird dropping reaches 33.4 C which is more than the hot area created by
the junction boxes. Factors like this will aect the shading loss of the array, as explained in
Section 5.5.4, which can only be improved by physically washing the dirty modules or waiting
for rain to do the job. It is, however, not given that rain will wash bird droppings since this can
be quite sticky and is best removed by physical labour.
There are 23 points in the array where string cables are connected together to form a string
of series connected PV modules. If these connectors are not connected properly, heat will
evolve, which will increase the voltage drop and form thermal heating at these points. Too
high temperatures will aect the cables and could, at worst case, cause a fire. Figure 5.20
88
Figure 5.20: Thermal inspection of the cable MC3 connectors. It is seen that no heat is
evolved at this point.
is illustrating one of these connection points. All of them were inspected with the thermal
camera, but no thermal heating was observed. Figure 5.20 shows that the connectors are
properly connected and does not show any abnormal temperatures.
5.8
Inverter Performance
With the included software it was only possible to extract 15 min data from the inverter. With a
Python script received from Eltek and modified by our supervisor, we were able to extract onesecond data from the web interface. Due to time limitations, only a few days of logged one-second
data are analysed in this thesis. The data was used to analyse how the inverter is handling low
irradiance and overirradiance conditions, in addition to calculating the performance ratio. The
energy harvest was recorded from the Theia Analyzer and is based on 15 min data, which
covers approximately one month. Screenshots of the Theia Analyzer pages and the Theia
Service tool from where the inverter data was downloaded is given in Appendix I.
5.8.1
From analysing Figure 5.21, two colours can be seen where the green line (PAC ) is almost
consistently under the blue line (PP V ). The PP V peak at almost 14:00 hours is recorded to
be 2.065 kW, and the respective PAC value is recorded to be 2.006 kW, illustrating that the
inverter works at its full capacity and efficiency close to the expected PV output on days with
close to STC conditions and little losses (Section 5.4.1).
89
(b) Irradiance measured over Grimstad with a mc-Si Cell and pyranometer.
Before and after the inverter starts generating power there are several blue spikes from the PV
array in Figure 5.21(a) and 5.27(a), of both DC current and power which lasts from approximately 6 to 11 minutes. These are generated between 05:16 and 05:27 in the morning and in
the evening between 21:07 and 21:30, where they can range up to 300 W but are usually close
to 200 W. Figure 5.21(b) shows that the pyranometer and mc-Si cell start recording at 6:30
in the morning and stop at 20:00. The peaks in Figure 5.21(a) occur earlier and later than
this, it is assumed that this is diuse or low irradiance on the PV array because of the early
sunrise at approximately 4:30. The blue peaks are therefore power from the PV system not
90
Figure 5.22 show the first minutes when the PV voltage is under the minimum inverter PV
voltage, between 5:24 and 5:36. Between 5:36 and 5:42, the inverter MPPT with a minimum
voltage of 230 VDC is searching for the array MPP, it is found at approximately 5:42 when the
inverter is receiving the required 7 WDC and Figure 5.21(a) start indicating that the inverter is
producing power. The straight voltage line and spike in Figure 5.22(b) at approximately 5:30
occurs after the diuse power spike in Figure 5.27(a) and indicate how the inverter is receiving
an increased input voltage as it is searching for the required input power and MPP. From
Figure 5.22(a) it is seen how the input voltage is dependent on temperature in Figure 5.23. In
the morning, when the temperature is low, the voltage is high and reaches its morning peak at
approximately 8:15 and 11.2 C. It then decreases throughout the day as the temperature rises,
91
Figure 5.23: Module Temperature on 2 May 2014. 459, Titan and New mc-Si are
dierent measuring points on the array.
and a new voltage peak appears at 18:00 when the temperature is approximately 14.4 C. The
voltage then decreases as the sun sets.
As the inverter start feeding power into the grid, the current in Figure 5.24(b) is not consistent.
IAC is oscillating between 0.06 A and 0.4 A for almost an hour, even though it can be seen that
IP V has a stable and close to linear increase. This oscillation can be seen more clearly on cloudy
days when the array is exposed to low irradiance. When analysing Figure 5.25, it can be seen
that oscillation starts when irradiance is low. This is seen in the morning and evening when the
incident angle of the sun is low or the inverter is operating on diuse irradiance. The efficiency
during this time is illustrated in Figure 5.27(b). It is seen that the efficiency is not stable, but
is increasing as the inverter is starting up and operates at better efficiencies. Figure 5.25(b)
indicates that the inverter output does not only oscillate at these times, but also at times when
Figure 5.25(c) indicate low irradiance over a long period of time, like at midday between 12:00
and 14:00. It would be interesting to investigate how oscillation over a long period of time,
like in Figure 5.25(b) aects the power quality, but this is not investigated in this thesis. This
oscillation could however, introduce harmonic currents and aect the power quality of the grid.
It is required to keep these harmonics as low as possible or at least keep it within the limits
stated in Tables 5.4 and 5.5.
The Theia HE-t datasheet in Appendix A explains that Early start-up and high efficiency at
low irradiation gives longer operation time and higher energy yields. Figure 5.21(a) indicates
92
Figure 5.24: Inverter current with the oscillation at low irradiance on 2 May 2014.
a high power at high irradiance, and early start-up which results in longer operation time is
consistent with the figures. As already explained, an early start-up at low irradiance results in
more oscillation, which could aect the output power quality. Figure 5.25(a) indicates that the
inverter has a good efficiency at low irradiance, but the efficiency is not consistent. It varies
a lot, and drops as far down as 43 % at 12:35 and 22 % at 16:45 for conditions where Figure
5.25(c) indicate low irradiance. These drops can last from 5 minutes to one hour, dependent
of the irradiance, before the inverter manage to increase the efficiency. This is also illustrated
in Figure 5.26 where PAC,rms is at some times much lower than PP V . Eltek has been informed
about these results, but it is assumed that they are already aware of the oscillation eect at low
93
100
[%]
80
60
40
20
0
5
10
15
20
Time [ h ]
IPV
IAC,rms
Current [ A ]
0.8
0.6
0.4
0.2
0
10
12
14
16
Time [ h ]
18
20
180
GHI
mcSi Cell
Pyranometer
Mean G
160
Irradiance [ W/m2 ]
140
SELF
120
100
80
60
40
20
10
12
Time [ h ]
14
16
18
irradiance.
Figure 5.27 illustrates a phenomenon appearing every day. From 06:53 to 7:11, the efficiency
drops suddenly despite of an almost exponential increase in input current (Figure 5.24(b) and
an almost constant increase and less varying (i.e more dense) input voltage (Figure 5.22(b)).
The drop lasts for 18 minutes or until the inverter manages to increase it back to the original
94
course. The efficiency then reaches 96 % before it increases to the maximal efficiency of 98 %,
when reaching this point it stop being stable, drops a few percentage and start varying. This
drop in efficiency results in the power loss illustrated in Figure 5.27(a) of only 14 W to 29 W,
a rather insignificant power loss appearing every morning and evening. The time range varies
from 6:48 to 07:35 in the morning and 18:11 to 20:22 in the evening, for the period 2-6 May. It
is consistent and is easiest to detect on clear sky mornings and evenings when the power drop
can be easily read o the power curve, like in Figure 5.27(a). Even though a power drop can
be easily detected by studying the efficiency curve in Figure 5.27(b), it is harder to detect on
cloudy days like Figure 5.26 illustrates. By analysing Figures 5.25(a) and 5.25(b), it can be
seen that the large drops in efficiency happens in a transition when the current stop oscillating
and start producing a higher AC current. Since the input current and voltage is relative stable,
i.e. not oscillating like the output current, it is believed that external influences, software or
hardware programming or an algorithm design causes this.
The current efficiency curve in Figure 5.28(a) is studied for two dierent days, 2 May and 18
May. It was recorded a clear sunny day on 2 May with some thin clouds in the afternoon (Figure
5.21(b)), while 18 May was a more overcast day with varying sky conditions as illustrated in
Figure 5.29. Figure 5.28(a) illustrates the efficiency versus input current on these days, showing
a consistent drop in efficiency at currents between 0.3 A and 0.4 A and another consistent drop
at 98 % efficiency and 0.8 A. The figure shows a more varying and less consistent efficiency for
the clear day (2 May) between 0.8 A and 2 A, where the efficiency on the cloudy day is a lot
more stable at this range. This changes above 2 A where the inverter manages to operate at
a more stable efficiency. Every dot in Figure 5.28(a) illustrates one second of data, it is seen
95
Figure 5.27: A drop in power, illustrated with the efficiency on 2 May 2014.
that for a few seconds the efficiency drops down to 89 and 90 % and then it increases up to
98.5 %. An imaginary line can be interpolated between the steady efficiency of 98 % at 0.8 A
and the maximum efficiencies, creating a theoretical maximum efficiency of 98.5 %. This is an
efficiency only achieved for a few seconds at a time, but they form a line that is more prominent
at currents between 3.5 A and 6 A. Since no specific pattern can be seen between the available
days of measurements, no conclusion can be made of why the efficiency is more stable on a
cloudy day at low currents and at sunny days at high currents. It is believed that this is a
result of the varying irradiance on dierent days determining how the MPPT should operate.
The MPPT is not perfect, Figure 5.28 shows how the MPPT is having difficulties to follow the
MPP of the array as the current and irradiance changes. The I-V curve from Figure 5.28(b)
shows the relation of the input voltage and current. It is seen that this curve have the same
96
(b) I-V curve from a clear day (2 May) and a cloudy day (18
May).
Figure 5.28: Input current and voltage are compared on two days with dierent sky conditions.
Irradiance [ kW/m2 ]
1.5
PV Cell
Pyranometer
0.5
0
6
10
12
14
16
18
20
Time [ h ]
Figure 5.29: Irradiance on 18 May 2014, indicating varying irradiance at midday and in the
afternoon [19].
tendencies as Figure 5.28(a), the voltage is varying at low currents on a sunny day and varying
at high currents on cloudy days. It is also seen that at currents between 0.8 A and 2 A, the
97
voltage is high while the current efficiency drops. The efficiency in Figure 5.28(a) then increases
at higher currents and slightly decreasing voltages.
Figure 5.31: Overirradiance over Grimstad of almost 1.5 suns on 9 May 2014, the time values
are approximate [19].
Overirradiance is a phenomenon aecting the output of the solar array and occur when clouds
are boosting the existing irradiance to higher values. The irradiance increases when the spread
of light is being magnified in optically thin clouds, like Figure 5.30 illustrates. These bursts of
overirradiance is dependent of clouds and could last from a few seconds to minutes at a time
[94]. In a climate with a high rate of clouds, it would be common and likely for the PV array
to be exposed to overirradiance. It would therefore be of interest to analyse how the inverter
works at these conditions, which occurred on 9 May. Figure 5.31 illustrates the overirradiance
reaching 1.48 suns (1485 W/m2 ) and lasts for approximately 30 seconds. During this time, the
98
Figure 5.32: Power and current produced during overirradiance on 9 May 2014 [19].
inverter records an input power of 2.9 kWDC and input current of 8 ADC while it produced a
peak power of 2.8 kWAC and a peak current of 11.8 AAC , which is shown in Figures 5.32(b)
and 5.32(a). The temperature on the back of the PV array at this time was 20-22 C, the
equivalent cell temperature is a few degrees higher than this and is assumed to be close to STC
temperature. Meaning that the increase in power was caused by overirradiance and not a result
of temperature eect on the PV power. These figures show how fast the inverter respond to
sudden irradiance changes. No delay in increasing or decreasing power or current is noticeable,
even small changes are almost identical between IP V and IAC in Figure 5.32(a) which allows
the inverter to convert as much of the PV input as possible.
Figure 5.33(a) shows the maximum efficiency of 97 % achieved at the time of the overirradiance,
99
for a few seconds the efficiency even reaches 98.5 %. Even though the irradiance varies, the
efficiency remains at a relatively stable value between 92 % and 97 % with an overall good
efficiency. Better efficiencies at high currents are illustrated in Figure 5.33(b) showing that at
almost 1.5 suns and an input current of 8 ADC , an efficiency of 97 % is reached. The figure
indicates that during this time, at a current between 7 and 8 A, the efficiency is much more
consistent than at lower currents when the irradiance is also lower. Since this only happened
for a few seconds, it shows that the MPPT operate better at higher input currents and higher
irradiance when the efficiency is varying less.
100
(b) Inverter efficiency for SMA 2500 inverter. Measured in Klamath Falls, Oregon in June 2006 [95].
Figure 5.34: Inverter efficiency of Theia HE-t compared to SMA 2500 on a sunny day in
Grimstad and Klamath Falls, respectively.
Despite a few efficiency drops at low irradiance, the inverter has a good overall efficiency. Figure
5.34 compares the efficiency of Theia HE-t on 2 May 2014 to the SMA 2500 in June 2006. It
shows how the efficiency of the SMA 2500 decreases with 2-3 % during midday at an average
ambient temperature of 26 C , compared to the efficiency of THEIA HE-t over the same time
with a module temperature of 33 C at 14:00 [95]. The temperatures for the curves in Figure
5.34 are dierent, it is therefore hard to compare the two efficiency curves when the conditions
are not the same. It is, however, observed that the module temperature does also increase in
Grimstad from 11 C in the morning to 33 C at midday. While [95] explains how the SMA 2500
efficiency decreases as the solar radiation is increasing above the inverters optimum operating
point of 400
600 W/m2 , the Theia HE-t manage to hold a relative stable efficiency varying
within 1 % at the same time. The SMA 2500 does however; manage to hold a more stable
101
Discussion
As mentioned, the inverter has a good overall efficiency varying at approximately 97 % measured
on a sunny day as illustrated in Figure 5.34(a). It has a good performance, producing close
to the rated power of 2.116 kWp on days close to one sun and up to 2.8 kW on a day with
overirradiance close to 1.5 sun.
Figures 5.27(b), 5.28(a) and 5.34(a) illustrate a phenomenon appearing every day under conditions where the input voltage is increasing and input current has an almost exponential increase
without any peaks or drops. This phenomenon is only seen on the output current and power
as well as on the efficiency curves and was therefore believed to be a result of a hardware issue,
external influences or how the inverter algorithm is programmed. This suspicion was confirmed
by Eltek [82]. At a power smaller than 200-300 W the Theia HE-t uses dierent algorithms
for the inverter and the DC/DC-converter to ensure a high efficiency. When analysing Figure
5.27(b), this change of algorithm is seen by the very evident drop down to 70 %. The increase in
efficiency before this drop is the result of one algorithm while the drop marks the spot where the
inverter and DC/DC converter is changing to a dierent algorithm. It is seen that the inverter
switches algorithm at 7:30, at a PV power of 77 W. This happens because the inverter has a
hysteresis, which is dependent of irradiation and voltage. The inverter maintain the increase
in efficiency until it reaches a point decided by the algorithm (77 W in Figure 5.27(a)) where
the inverter start operating at normal mode and the efficiency drops. The variable efficiency
occurring after 7:30 in Figure 5.27(b) is therefore a result of the inverter operating with another
algorithm at normal mode.
This change of algorithm is also seen in Figure 5.28(a) and is believed to be the reason for
the very evident drop in efficiency between 0.3 A and 0.4 A, a delay is then present when the
inverter operates at an increasing efficiency between 0.4 A and 0.8 A before it indicates when
the inverter is running at normal mode between 0.8 A and 2 A. The long efficiency drop in
Figure 5.33(a) is also a result of that. The drop lasts for almost two hours because the irradiance
decreased after the inverter switched algorithms and the efficiency was no longer optimal. If
the irradiation had dropped more, then the inverter would have switched back to the other
algorithm.
102
It is believed that the oscillation in Figure 5.24(b) is also a result of a special algorithm being
used. It seems like one algorithm is programmed to reduce losses related to the inverter switches
at low irradiance and power to maintain a good efficiency at these conditions. This algorithm
is especially used on days with low irradiance like on 6 May in Figure 5.25(b) where some of
the large drops in efficiency happens almost simultaneously to when the current start and stop
oscillating. It is therefore assumed that the oscillation is a deliberate result of an algorithm
designed to maintain a good efficiency at low irradiance, and that the resulting harmonics is
within the requirements stated in IEC 61000-3-2 and IEC 61000-3-12.
During short bursts of overirradiance, both the current and power exceeds the expected power of
2.1 kWp calculated in Chapter 5.4.1 and cable CCC of 7.81 A calculated in Chapter 5.4.2. This
power is calculated for conditions at STC (1000 W/m2 ), which is the average peak irradiance
at local conditions [94]. As illustrated in Figure 5.32, both current and power is higher because
of new conditions with a higher irradiance. It is therefore important to account for these
overirradiance conditions when designing the PV system. The inverter must be large enough to
handle overirradiance and to avoid loosing this extra power.
Both the Australian [3] and German manuals [4] state that the cable CCC should be calculated
by using respectively 1.2 and 1.25 times array ISC . The CCC for this PV system, calculated
with guidelines from [4], is lower than the 8 ADC produced during overirradiance. It indicates
that if oversized cables were not used, the cable could be damaged over time as a result of
repeatedly overirradiances. This could happen if more current than the CCC runs through the
cable and is avoided by adding a circuit breaker. The circuit breaker would, however, have to
be manually re-set for the inverter to start producing again. Some time could pass before this
is done where the inverter would not be operating.
The Eaton circuit breaker is made for PV systems and has settings to allow for overcurrent to
run through it for a short period of time and then allows the inverter to utilize the overirradiance.
Overirradiance occurs mainly in the end of May, June and July and can happen quite often and
rapidly, depending on the clouds. Since cables with larger cross section can only reduce the cable
loss, the Australian and German manuals are considered to be too conservative when calculating
the CCC and the cable cross section. Neither of these manuals mention overirradiance or how
to handle this.
5.8.2
103
Performance Ratio
As mentioned in Section 5.5.11, performance ratio measures the quality of the PV system and
describes the relation between the theoretical energy and the real energy output from the PV
system. Usually one year of data is needed to perform an optimum PR analysis, but this is
not possible to achieve in this thesis, and therefore only five days of data were used. Equation
(5.56) was used to calculate the PR values [96].
PR =
E
ArH
(5.56)
M ax P V module power
P V module area
0.092 kW
0.84 m2
= 11 %.
Table 5.17: Daily and total irradiation; electrical energy from inverter to grid and PR based
on five days of data.
Date
02.05.14
03.05.14
04.05.14
05.05.14
06.05.14
Total
PR [%]
92
91
91
67
72
89
From Table 5.17 it can be seen that the PR was particular high at sunny days with high
irradiation levels. The PR is lowest at 5 and 6 May when the irradiation is lower. However, the
PR is higher on 6 May even if the irradiation is lower when compared with 5 May. Figure 5.35
shows the temperatures of 5 and 6 May. On 5 May the average array temperature was 11.56 C
with a average peak temperature of 19.3 C, and on the 6 May the average temperature was
9.24 C with a average peak temperature of 10.92 C. In addition the temperature was for a
longer period > 1 hour higher than 14 C on May 5. The PV array is more efficient at lower
temperatures and this could explain why the PV system was able to harvest more energy even
though the irradiance was lower on the 6 May.
The total PR from five days of data is 89 %. This means that 11 % during this period were not
converted to usable energy due to system losses. There are many factors that could influence
the PR value. First of all the measurement period is too short and this gives insufficient
measurements for calculating a reliable PR. After one year the system losses would probably be
22
20
104
459
Titan
New mcSi
18
TMOD [C]
16
14
12
10
8
6
4
6
10
12
Time [h]
14
16
18
20
(a) Module Temperature on 5 May 2014. 459, Titan and New mc-Si are dierent measuring
points on the array.
11.5
11
459
Titan
New mcSi
10.5
TMOD [C]
10
9.5
9
8.5
8
7.5
7
6
10
12
Time [h]
14
16
18
(b) Module Temperature on 6 May 2014. 459, Titan and New mc-Si are dierent measuring
points on the array.
closer to 20.4 %, which was calculated in Section 5.5.10. Temperature variations and temporary
shading could have a larger eect on the calculated values if the measurement period is only a few
days. At least one month of data should have been included in the calculations to minimise the
eect from these factors. In addition to temperature and shading, conduction losses, efficiency
factor of the PV modules and the inverter also aects the PR values [97].
5.8.3
Energy Harvest
From 24 April, when the inverter was installed and started commissioning to 26 May, the inverter
has been operating for 493 hours. When using Equation 5.37 and the average irradiation for
105
kW h
,
m2
calculated with no losses added. When assuming the voltage drop, inverter efficiency, relative
efficiency, spectral gain, angular reflectance eects and dirt losses from Section 5.5, and assuming
no losses due to shading, tilt angle and manufacturers tolerance, the estimated power in May
is calculated to be 289 kWh. The expected power output will in reality be slightly smaller
when considering that the inverter only operated for six days in April and 26 of 31 days in May
(months with dierent average irradiation), an uncertainty of 5 % can also be added due to
uncertainty in the system losses and varying temperature and irradiance during the operating
hours. This varying irradiance is illustrated with Figure 5.36, showing the power output during
the last week of operation. As seen on 19, 21, 22 and 23 May, clouds and varying irradiance
varied the daily power production which resulted in lower power output on these days compared
to 24 May.
Figure 5.36: A screenshot of Theia Analyzer showing the inverter power output during the
last 7 days, from 19 to 25 May. It indicates a varying power output due to irradiance, a very
low production on 21 May and a production of 1828 W on 24 May.
The total produced power so far on 26 May, the month of May and year 2014 is showed in
Figure 5.37. The total energy produced in April and May in Figure 5.37 c) is assumed to be
one month to simplify the calculations above, resulting in an expected production in May
of 289 kWh. It is seen in Figure 5.37 and Appendix I that the inverter produces more power
than expected in the calculations, meaning that either the irradiation in May is higher than the
average irradiation of the past three years, or the system losses at this period is smaller than
expected.
106
Figure 5.37: A screenshot of Theia Analyzer showing a) The total energy produced until
13:00 on 26 May of 8661 Wh, b) Energy produced every day in May, giving a total energy of
238 kWh this month, c) Energy produced so far this year, 313 kWh has been produced since
the inverter started commissioning.
It is seen from Figure 5.37 b) that the inverter has a good average daily production, except
from a few days with little or close to no production, like 6, 8 and 11 May. May was a sunny
month, as Figure 5.37 b) indicate, and the inverter efficiency has been operating in average very
close to its maximum capacity. This resulted in a good utilisation of the available irradiation
on days with and without clouds. The 313 kWh produced in April/May is more than what is
expected with the given assumptions in May, Figure 5.37 therefore shows a good production rate
of the inverter on a daily, monthly and yearly basis. It is assumed that this good progression
will continue to increase throughout May and the summer months as the irradiation increases,
until it is expected to decrease in September. Assuming if the trend found by the production in
May, shown in Figure 5.37 c), related to the irradiation in Table 4.3 is maintained throughout
a whole year, the progression looks good with regards of producing the expected yearly energy
yield calculated in Equation 5.39 of 2088 kWh or more. It is however, hard to predict a trend
from Figure 5.37 c) and to conclude that the system will be operating accordingly when only
one month is used as a reference. Ideally, two or three additional months should have been
analysed to indicate a clear increase or decrease in the monthly production trend line.
Chapter 6
Conclusion
Renewable energy sources are now being implemented in a large scale in Europe. Large and
small-scale hydro power plants, wind turbines and PV systems are built to reach the EU goal
of getting 20 % of its energy from renewable sources by 2020. Energy from Grid Connected
PV Systems is one way of reaching this goal when considering that only 0.01 % of the solar
energy reaching the earths surface is needed to cover the human energy need. As PV systems
are covering more of the electricity demand on earth, at least 36.9 GW of them was installed
all over the world in 2013. Europe, which has been a global leader on the PV market for
many years, installed only 10.3 GW. This was a decrease of 6.7 GW from 2012. Germany,
being number one in the European PV market, installed 3.3 GW of the European PV systems
while Norway installed only 0.6 MW in 2013. When the average monthly irradiation found in
Hamburg of 91.8 kW h/m2 is compared to the average irradiation in Grimstad of 86.6 kW h/m2 ,
using PVGIS, it can be seen that the solar energy possibilities in the southern parts of Norway
is worth investigating for future installations.
To examine these possibilities, a PV system was built at the University of Agder in Grimstad
while making sure that the power produced and system components meets the required demands
from the local utility company, Agder Energi. Theia HE-t 3.8, a string inverter from Eltek was
found to be the most reasonable and compatible inverter, fulfilling all the power quality and
protection requirements given from Agder Energi in addition to having good logging capabilities.
It was calculated that the inverter could be connected to one string of 23 de-rated NAPS
(NESTE) NP100G12 PV modules. The University provided 48 PV modules, which were tested
with regards to Imp , Vmp and Pmax to find the best IV-curve characteristics so that the best 23
107
Chapter 6 Conclusion
108
modules with the most similar and highest current ratings were used to constitute the system
of 2.116 kWp .
The PV array is made out of 0.83 m2 sized modules, and should have to withstand a wind
load of 1.26 kN each. SunTech modules available on the market can withstand wind loads of
at least 2.4 kN/m2 . There is no available data for the NAPS (NESTE) modules used in this
PV system, they are old and it is therefore unknown if they can withstand the same load as
SunTech. Since the maximum expected wind load is much lower than the minimum wind load
sustainable by the more modern PV modules, it was assumed that the older PV modules are
capable of withstanding this maximum wind load.
For the practical examination of the solar energy possibilities, the PV system with an estimated
total system loss of 20.4 % was installed and expected to produce 2088 kWh/year with a
performance ratio (PR) of 80 %. Start-up of the PV system started on 24 April, when we got
permission from Agder Energi to finalize the remaining agreements after the PV system had
started commissioning. Data from the following days of operation were analysed to see how the
inverter operated concerning efficiency. The first days of clear sky indicated that the inverter
was operating at a good efficiency of 97 % and producing a maximum power of 2.065 kWp at
irradiance close to one sun. The inverter was operating close to its maximum efficiency and
responded almost instantaneously to sudden changes in the irradiance.
It was observed that the inverter had problems producing a steady current when the system
was exposed to very low irradiance at midday and at early operating hours. The current curves
show a consistent oscillation at times with very low irradiance and a clear drop in efficiencyand power curves approximately at the same time in the morning and evening. It is also
observed that when the current oscillates on days with low irradiance, the efficiency drops in
the transition between oscillation and a steady current. Eltek was aware of these observations,
which was caused by the inverter using two dierent algorithms for low- and high irradiance to
ensure a good efficiency at these conditions. The inverter switches algorithms at a power less
than 200-300 W which will cause a drop in efficiency at that time. The algorithms are operating
with hysteresis, which is dependent of irradiance and voltage. The length of this efficiency drop
therefore depends on the available PV power and could vary from 10 minutes to two hours.
One day of extreme overirradiance occurred during the time period, which resulted in a maximum peak power of 2.8 W at almost 1.5 suns and cell temperature of about 24 C. During the
few seconds of overirradiance, the inverter responded quickly and the MPPT managed to follow
Chapter 6 Conclusion
109
the MPP of the array to a good extent. DC Current of 8 A was observed at a short time while
the inverter kept operating at maximum efficiency, which indicates that the inverter manages
to fully utilize the power from overirradiance conditions. However, this happens because the
system is oversized to handle currents higher than what is possible for the system to achieve. If
the system was not oversized, the circuit breaker would have dierent tripping times at currents
with higher values than the rated. To avoid this, and to make sure that power from overirradiance is fully processed, it is recommended to size system cables and circuit breakers for the
maximum overirradiance conditions of 1.6 suns in Grimstad.
During the time period of this thesis, the PV system was operational for 6 days in April and 26
days in May which was, for the sake of calculations, assumed to be approximately one month.
The system produced 313 kWh during this time, which was more than the expected power of
288 kWh calculated for May. A PR of 89 % was calculated for the system during the first
week of May. It was seen that the PV system has a good PR of 92 % on sunny days with high
irradiation and a varying smaller PR on days with lower irradiation. It has good preconditions
to meet the expected yearly production of 2088 kWh or more. However, it is hard to predict a
trend line for how the system operates in the future when only one month is used as a reference.
For further work, we recommend the PV system to be operational for minimum one year. Several
months of data makes it possible to compare energy output with estimations from tools such
as PVGIS. In addition, it would be beneficial to see how much power a small sized PV system
can generate throughout a year in Southern Norway.
Overirradiance and the resulting power gain should be studied. Furthermore, it should be
determined how much harmonics are caused as a result of oscillation in the current when the
inverter operates on the low irradiance algorithm. It would also be interesting to see how
the power quality changes throughout the day. More studies should be done regarding the
albedo eect. More research need to be done to see how the sea in the horizon and snow at
the installation site is contributing with diuse irradiance and how it aects the PV system.
Another factor worthy of study is the shading eect on the PV array. A more precise loss factor
due to shading can be determined if more research is done to see how shading from the telecom
tower and fence aects the array.
Bibliography
[1] European Commision. Renewable energy. http://ec.europa.eu/energy/renewables/
index_en.htm, Last accessed January 2014.
[2] IEA International Energy Agency. PVPS Report snapshot of Global PV 1992-2013 - Preliminary Trends Information from the IEA PVPS Programme. http://www.iea-pvps.
org/index.php?id=3&eID=dam_frontend_push&docID=1924, Last accessed May 2014.
[3] G. Stapleton, S. Garrett, S. Neill, and B. McLean. Grid-Connected PV Systems Design and
Installation. Global Sustainable Energy Solutions (GSES), 7th edition, December 2010.
[4] The German Solar Energy Society. Planning and Installing Photovoltaic System. A guide
for Designers, Installers, Architects and Engineers. Routhledge, 2013.
[5] A. Ashfaq. Power Electronics for Technology. Prentice Hall, 1st edition, June 1998.
[6] Y. Chen and K. M. Smedley. A cost-eective single-stage inverter with maximum power
point tracking. IEEE Transactions on Power Electronics, 19(5):12891294, September
2004.
[7] S.B. Kjaer, J.K. Pedersen, and F. Blaabjerg. A review of single-phase grid-connected
inverters for photovoltaic modules. IEEE Transactions on Industry Applications, 41(5):
12921306, September 2005.
[8] S. Bergsmark. UiA. Private Communication, 2014.
[9] S. J. Chapman.
http://www.mhhe.com/engcs/
http://www.fronius.com/cps/rde/xbcr/SID-17106601-1811CF4A/
110
Bibliography
111
fronius_international/SE_TA_High_Frequency_Transformer_With_Transformer_
Switchover_EN_320487_snapshot.pdf, Last accessed February 2014.
[11] Schneider Electric. Description of Galvanic Isolation. http://www.schneider-electric.
us/sites/us/en/support/faq/faq_main.page?page=content&country=US&lang=
en&id=FA157465&redirect=true, Last accessed February 2014.
[12] V. W. Quinn. Scaling Design Methods To Improve Transformers. http://www.dalitech.
com/Resources/Scaling%20Design%20Methods%20to%20Improve%20Transformers.pdf,
Last accessed May 2014.
[13] E.
Seymour.
ers.
Design
Characteristics
of
High-Power
Photovoltaic
Invert-
http://solarenergy.advanced-energy.com/upload/File/White_Papers/
Chen.
ers.
Integration
of
Isolation
for
Grid-Tied
Photovoltaic
Invert-
http://www.analog.com/static/imported-files/tech_articles/
Integration-of-Isolation-for-Grid-Tied-Photovoltaic-Inverters-MS-2356.pdf,
Last accessed February 2014.
[15] D. Chiras. Solar Electricity Basics: A Green Energy Guide. New Society Publishers, July
2010.
[16] J. Worden and M. Z.-Martinson. How Inverters Work. http://pvdepot.com/media/
resources/How-Solar-Inverters-Work.pdf, Last accessed March 2014.
[17] Y.-C-Kuo and T.-J. Liang. Novel maximum-power-point-tracking controller for photovoltaic energy conversion system. IEEE Transactions on Industry Electronics, 48(3):594
601, June 2001.
[18] G. Hering and C. Warren. Harvest time. Photon International - The Solar Power Magazine,
(10), 2011.
[19] G. H. Yordanov. UiA. Private Communication, 2014.
[20] NVE. H
andtering av plusskunder og vedtak om dispensasjon fra forskrift 302 om konomisk
og teknisk rapportering m.v. http://www.nve.no/PageFiles/9733/Plusskunder.pdf,
Last accessed March 2014.
[21] R. H. Josefsen. Agder Energi Nett. Private Communication, 2014.
Bibliography
112
[22] Dr J. Ravishankar. UNSW. Types of earthing systems. Lecture notes from the course
ELEC9716 - Electrical Safety, 2013.
[23] Schneider
Electric.
Photovoltaic
system:
electrical
equipments
selection.
http://www.electrical-installation.org/enwiki/Photovoltaic_system:
_electrical_equipments_selection, Last accessed March 2014.
[24] Hafslund.
http:
//nettbiblioteket.hafslundnett.no/NettLib/Retningslinjer/01%20Tilknytning/
06Tilknytning%20av%20produksjonsanlegg/01_06_01_Produksjon_230_400V.pdf,
Last accessed March 2014.
[25] H. T. Bjrnstad.
Bibliography
113
[32] G. H. Yordanov. Characterization and Analysis of Photovoltaic Modules and the Solar Resource Based on In-Situ Measurements in Southern Norway. PhD thesis, NTNU, November
2012.
[33] A. Christensen. Lys framtid for solkraft i nord. http://forskning.no/artikler/2012/
november/338931/, Last accessed March 2014.
[34] European Commision. Photovoltaic Geographic Information System. http://re.jrc.ec.
europa.eu/pvgis/, Last accessed May 2014.
[35] D. Heinemann. Energy Meteorology. http://www.energiemeteorologie.de/team/dehe/
Documents/Energy_Meteorology__%20Script.pdf, Last accessed February 2014.
[36] L. Wald M. Lefevre, M. Albuisson. Description of the software Heliosat-II for the conversion
of images acquired by Meteosat satellites in the visible band into maps of solar radiation
available at ground level. http://www.soda-is.com/heliosat/heliosat2_soft_descr.
pdf, Last accessed March 2014.
[37] Kjeller Vindteknikk. Resource mapping of solar energy - an overview of available data
in Norway.
http://www.solenergi.no/wp-content/uploads/2010/01/KVT_OB_2013_
http://www.plgpower.com/
Bibliography
114
Hent klimadata.
http://lmt.bioforsk.no/agrometbase/
sion 20140101.
http://file4.fanyacdn.com/imglibs/files/2014%20Installation%
http://www.homepower.com/articles/
Weliczko.
Galvanic
Corrosion
Considerations
for
PV
Arrays.
http://solarprofessional.com/articles/products-equipment/racking/
galvanic-corrosion-considerations-for-pv-arrays?v=disable_pagination,
Last
Bibliography
115
http://new.abb.com/docs/librariesprovider22/
technical-documentation/solar-inverter-pvs300-flyer.pdf?sfvrsn=2,
Last ac-
in
The
a
Danfoss
user-friendly
DLX
UL
design.
inverter
series
Performance
and
flexi-
http://www.danfoss.com/nr/rdonlyres/
Bibliography
116
3769ba0a-9c73-455c-bb15-9e878b48139a/0/web_dksipfp203a122_dlx_factsheet_
us.pdf, Last accessed February 2014.
[66] S. Madsen. Soltek. Private Communication, 2014.
[67] GETEK. Fronius IG Plus. http://www.getek.no/brosjyrer/FroniusIG_plus.pdf, Last
accessed January 2014.
[68] Fronius International GmbH.
Fronius IG Plus.
http://www.fronius.com/cps/rde/
xbcr/SID-C82B8BB3-7A0C9337/fronius_international/DBL_Fronius_IG_Plus_M_06_
0008_EN_513_as13_low_156927_snapshot.pdf, Last accessed January 2014.
[69] Fronius International GmbH. Fronius Datamanager. http://www.fronius.com/cps/rde/
xchg/fronius_international/hs.xsl/83_27954_ENG_HTML.htm#.Uym449xyfbM,
Last
http://lovdata.no/
Planleggingsbok
for
kraftnett
http://www.ren.no/c/document_library/get_file?uuid=
Spenningskvalitet.
Bibliography
117
74190d68-f0bc-49a8-9fce-67eae2d73c1a&groupId=10206,
Last
accessed
February
2014.
[78] Selectricity.
Harmonic
Distortion
Analysis.
http://selectricity.com/
http://www.standard.no/no/Nettbutikk/
http://www.standard.no/no/Nettbutikk/
http://www.standard.no/no/Nettbutikk/
H
andbok Spenningskvalitet - Kortvarige underspenninger - Spenningsdipp.
http://www.sintef.no/Projectweb/Handbok_Spenningskvalitet/Kort-innforing/
Kortvarige-underspenninger---Spenningsdipp/, Last accessed March 2014.
[84] SINTEF.
H
andbok
Spenningskvalitet
Kortvarige
overspenninger.
http:
//www.sintef.no/Projectweb/Handbok_Spenningskvalitet/Kort-innforing/
Kortvarige-overspenninger-eng-swell/, Last accessed March 2014.
[85] J. E. Ormbostad. Montrh
andboka - NEK 400:2010, volume 1. Elforlaget, 2010.
[86] National Instruments. Part II - Photovoltaic cell I-V Characterization Theory and LabVIEW Analysis Code. http://www.ni.com/white-paper/7230/en/, Last accessed May
2014.
[87] R. A. Matula. Electrical Resistivity of Copper, Gold, Palladium and Silver. J. Phys. Chem.
Ref. Data, 8(4):11471298, 1979.
[88] T. M. Dauphnee and H. P.-Thomas. A Copper Resistance Temperature Scale. The Review
of Scientific Instruments, 25(9):884886, September 1954.
Bibliography
118
Fill factor.
http://pveducation.org/pvcdrom/
Appendix A
119
THEIA HE-t
The THEIA HE-t range defines a new level of
efficiency, flexibility and user friendliness for isolated
string inverters. Suitable for all PV cell technologies,
and ready for use all over the world, the THEIA HE-t
is the perfect choice for any PV installation.
THEIA HE-t
SOLAR INVERTERS: 2.0 kW 4.6 kW
PERFORMANCE
tMaximum efficiency 97.3 % with galvanic isolation
tSuitable for use with all PV modules of any technology, with the
ability to ground the positive or the negative terminal on the DC side
tCompliance with the highest international safety standards
tEarly startup and high efficiency at low irradiation gives longer
operation time and higher energy yields
EASE OF USE
tLightweight and easy to install
tWith or without DC Disconnect Switch
tColor screen with touch sense buttons
tIntuitive user interface
RELIABILITY
tHigh quality components, with a robust design
tBespoke Maximum Power Point Tracking
tStable operation under extremely dynamic irradiation conditions
tIP65 protection level
Eltek 2013
120
MODEL
2.0 HE-t
2.9 HE-t
3.8 HE-t
4.4 HE-t
4.6 HE-t
Nominal DC power
2100 W
3000 W
4000 W
4600 W
4800 W
Max. PV power
2625 Wp
3750 Wp
5000 Wp
5750 Wp
6000 Wp
Max. DC voltage
600 Vdc
9.5 A
13.5 A
18.0 A
21.0 A
21.0 A
Input features
INPUT DATA
OUTPUT DATA
Nominal output power
2000 W
2900 W
3800 W
4450 W
4600 W
2000 VA
2900 VA
3800 VA
4450 VA
4600 VA
Nominal AC current
9.0 A
13.0 A
17.0 A
19.5 A
20.0 A
Max. AC current
10.5 A
15.2 A
19.7 A
23.0 A
23.0 A
Mains frequency
50 Hz / 60 Hz (+/-10 %)
PERFORMANCE DATA
Maximum efficiency
97.2 %
97.2 %
97.2 %
97.3 %
97.3 %
CEC efficiency
96.8 %
96.8 %
97.0 %
97.0 %
97.0 %
EU efficiency
96.3 %
96.5 %
96.7 %
96.9 %
96.9 %
<7W
<1W
< 21 kg / 46 lbs
< 21 kg / 46 lbs
MECHANICAL DATA
Protection degree (EN 60529)
IP 65
Dimensions
Weight
< 19 kg / 42 lbs
Cable access
Bottom
< 19 kg / 42 lbs
< 21 kg / 46 lbs
DESIGN STANDARDS
EM compatibility
EN 61000-6-2, EN 61000-6-3
CE marking
Yes
Other standards
DIN VDE V 0126-1-1, G83/1, EN 50438, AS 4777, CEI 0-21, EN 61000-3-2, EN 61000-3-3, EN 61000-3-11,
EN 61000-3-12, IEC 62109-2, IEC 61727, UTE C 15-712-1, C10/11, VDE AR-N 4105, RD1663, G59/2
ENVIRONMENTAL DATA
Operating temperature
Storage temperature
- 30 C to + 80 C / - 22 to + 176 F
Ventilation
Convection cooling
ADDITIONAL FEATURES
Topology
I / III
Noise Emission
< 37 dB (A)
92
Communication
90
84
100
98
Warranty
96
94
88
86
82
80
78
76
74
72
70
Eltek 2013
121
Appendix B
122
Kristen Leifsen
Adresse:
Telefon:
918 73 575
E-post:
kristen.leifsen@uia.no
Produksjonsenhet:
Navn:
Grds- og bruksnr:
3/312
Anleggsadresse:
Jon Lilletunsvei 9
Postnummer:
4898
Kommune:
Grimstad
Status:
Tekniske data:
Type produksjon:
Solceller
Generatortype:
Antall generatorer:
3,8
3x10-3
3x10-4
Utbyggingskostnad [kr/kWh]:
6,00
Legg ved kartskisse, mlestokk 1:12500 for hele anlegget og 1:3125 for kraftstasjonen, se
f. eks. www.statkart.no
Sted:
Dato:
Grimstad
1.24.2014
Signatur:
Sknaden skal sendes til det nettselskapet som eier distribusjonsnettet der produksjonsenheten skal tilknyttes
123
124
125
126
Appendix C
Specifications of Inverters
127
!
Input!(DC)!
!
Max!PV!
MPP!
power!
voltage!
range!
PVS300!TL! 4500!Wp! 335M800!V!
300!
SMA!
4040!Wp! 250M480!V!
3800MUS!
600!V!
Max!
voltag
e!
900!V!
18!A!
18!A!
14.6!A!
Max!
current!
3800!W!
3800!W!
4000!W!
17!A!
16!A!
17.4!A!
Output!(AC)!
Nom!power! Nom!
current!
Yes!
Yes!
Trans
form
er!
No!
16.2!A!
18!A!
3500!W!
3800!W!
N/A!
17!A!
Yes!
Yes!
DLX!3.8!
3710!Wp! 230M500!V! 600!V!
THEIA!3.8!
HEMt!
Fronius!IG!
plus!35!
V/HZ!
!
180M276!V/!47M
63!Hz!
211M264!V/!
59.3M60.5!Hz!
230!V/!50!
Hz/60!Hz!
230!V/!45M55!
Hz!
230!V/!50!
Hz/60!Hz!
!
Euro!
effici
ency!
96.4!
%!
96!%!
96.7!
%!
96.7!
%!
95!%!
MC3,!MC4,!
Screw,!
Clamp!
Screw!
Screw!
MC4!
DC/AC!
connectors!
6!
3!
4!
4!
PV!
Communication!
strings!
3!
Screw!
Ethernet,!EIAM
485,!KNX!
Bluetooth!or!
RS485!is!
optional!
Ethernet,!WebM
server,!CAN,!
RS485!
Ethernet,!CAN,!
RS485!
WLAN,!
Ethernet,!
MODBUS!TCP,!
Datalogger,!
Webserver!
Weight!
27!kg!
38!kg!
21!kg!
23!kg!
23.8!kg!
128
Appendix D
129
DC
DC string circuit-breaker
DC string circuit-breaker PKZ-SOL 2-poles
Rated operational voltage 900 VDC
Rated current In 4, 7, 12, 20, and 30 A
For permissible string short-circuit currents Isc of 5 up to 22 A
Ie
Isc
Type designation
Article No.
4A
7A
12A
20A
30A
1,6-3A
2,6-5A
5-9A
9-15A
15-22A
PKZ-SOL4
PKZ-SOL7
PKZ-SOL12
PKZ-SOL20
PKZ-SOL30
144069
144120
120937
120938
120939
1
1
1
1
1
wa_sg05409
Connection diagrams:
PV string
Unearthed system
Earthed system
DC disconnection / PV inverter
15
130
DC
DC Switch-disconnector P-SOL
Field of application:
DC-Disconnection in photovoltaic systems between
PV Array and Inverter to switch off the energy
No polarity
Any mounting position
Spring work contacts
Tested according to IEC/EN 60947-3, UL508
Certificate TV-Rheinland
Application acc. to IEC 60364-7-712 and IEC 62548
Connection diagram
Unearthed system
Earthed system
Technical Data
P-SOL20
P-SOL30
P-SOL60
2
1000 VDC
20 A
500 A
700 A
DC-21 A
III
8 kV
1500
6m
2
1000 VDC
30 A
500 A
700 A
DC-21 A
III
8 kV
1500
5m
2
1000 VDC
63 A
1500 A
1500 A
DC-21 A
III
8 kV
1500
3m
58
93
76
265 g
35 mm
58
93
76
265 g
35 mm
IP20
2x (1-6)
18-10
1,7 Nm
-25C to +60C
Damp heat, constant
Damp heat, cyclic
2
100.000
<120
IP20
2x (1-6)
18-10
1,7 Nm
-25C to +60C
55
140
160
920 g
35 mm
2xM4x18
IP20
2x (1-35)
14-2
3 Nm
-25C to +60C
2
100.000
<120
100.000
<120
Electrical
Number of poles
Rated operational voltage Ue
Rated operational current Ie
Rated short-circuit making capacity Icm
Rated short-time withstand current 1sec. Icw
Utilization category
Overvoltage category
Rated impulse withstand voltage Uimp
Operating cycles electrical at Ue and Ie
Internal resistance
Mechanical
Width
Height
Depth
Weight
Mounting quick fastening on DIN rail acc. to IEC/EN 60517
Screw fastening
Degree of protection
Terminal capacity Flexible with end sleeve
mm
AWG
Tightening torque of terminal screws
Ambient temperature range
Climatic resistance acc. to IEC 60068-2-78
acc. to IEC 60068-2-30
Pollution degree2
Operating cycles mechanical
Operating cycles mechanical per hour
Dimensions (mm)
P-SOL20
P-SOL30
P-SOL60
16
131
DC
Connection diagram
Unearthed system
Earthed system
Technical Data
SOL20
SOL30
2
1000 VDC
20 A
500 A
700 A
DC-21 A
III
8 kV
1500
8m
2
1000 VDC
30 A
500 A
700 A
DC-21 A
III
8 kV
1500
5m
420 g
IP65
-25C to +60C
Damp heat, constant
Damp heat, cyclic
3
100.000
<120
420 g
IP65
-25C to +60C
Electrical
Number of poles
Rated operational voltage Ue
Rated operational current Ie
Rated short-circuit making capacity Icm
Rated short-time withstand current 1sec. Icw
Utilization category
Overvoltage category
Rated impulse withstand voltage Uimp
Operating cycles electrical at Ue and Ie
Internal resistance
Mechanical
Weight
Degree of protection
Ambient temperature range
Climatic resistance acc. to 60068-2-78
acc. to 60068-2-30
Pollution degree
Operating cycles mechanical
Operating cycles mechanical per hour
3
100.000
<120
Dimensions (mm)
SOL20(30)
a
[mm]
MC4
234
MV
224
SOL20/2MC4
SOL30/2MC4
SOL20
SOL30
Input
+ Strings
- Strings
SOL20/2MV
SOL30/2MV
Input (M12)
+ Strings
- Strings
SOL20/4MC4
SOL30/4MC4
Input
+ Strings
- Strings
Output (M16)
-
+
Output
17
132
DC
Connection diagram
Unearthed system
Technical Data
PKZ-SOL4
PKZ-SOL7
PKZ-SOL12
PKZ-SOL20
PKZ-SOL30
Electrical
Number of poles
Rated operational voltage Ue
Rated current In
Thermal tripping characteristic
Electromagnetc tripping characteristic
Rated ultimate short-circuit breaking capacity Icu
Rated service short-circuit breaking capacity Ics
Overvoltage category
Rated impulse withstand voltage Uimp
Operating cycles electrical at Ue and In
Internal resistance
2
900 VDC
4 / 7 / 12 / 20 / 30 A
1.05 to 1.3 x In
6 x In
5 kA
1.5 kA
III
8 kV
1500
138 / 60 / 32 / 14 / 9 m
Mechanical
Width
Height
Depth
Weight
Mounting quick fastening on DIN rail acc. to IEC/EN 60517
Screw fastening
Degree of protection
Terminal capacity Flexible with end sleeve mm
AWG
Tightening torque of terminal screws
Ambient temperature range
Climatic resistance acc. to IEC 60068-2-78
acc. to IEC 60068-2-30
Pollution degree
Operating cycles mechanical
Operating cycles mechanical per hour
58
93
76
265 g
35 mm
IP20
2x (1-6)
18-10
1,7 Nm
-25C to +60C
Damp heat, constant
Damp heat, cyclic
2
100.000
<120
Mounting position
PKZ-SOL4
PKZ-SOL7
PKZ-SOL12
PKZ-SOL20
PKZ-SOL30
90
90
90
18
133
Earthed system
DC
30
SettingEinstellwert
value at the
circuit breaker
am Schutzschalter
[A][A]
tripping time
PKZ-SOL30
25
20
PKZ-SOL20
15
PKZ-SOL12
10
PKZ-SOL7
PKZ-SOL4
0
0
10
15
20
25
Kurzschlussstrom
[A] [A]
Short-circuit
currentSolarmodul
solar module
1)
Dimensions (mm)
PKZ-SOL4
PKZ-SOL7
PKZ-SOL12
PKZ-SOL20
PKZ-SOL30
19
134
Appendix E
PV Module Ranking
135
0.146
0.041
0.077
0.080
0.052
0.116
0.120
0.103
0.132
0.068
0.235
0.104
0.078
0.057
0.129
0.033
0.094
0.102
0.105
0.105
0.054
0.034
0.083
0.0
0.1
0.1
0.0
#0.1
0.2
0.2
0.0
0.2
#0.1
0.0
0.0
#0.1
0.0
0.1
#0.1
0.0
#0.1
#0.2
0.0
0.9
0.1
0.0
0.1
0.3
#0.1
0.0
0.1
0.0
0.0
0.1
0.1
0.1
0.3
0.1
2.1
#0.1
0.1
#0.1
0.0
0.1
0.0
#0.1
#0.2
0.1
0.0
#0.1
0.3
#0.5
Pm
Pm#Im*Vm
Median StdDev
15.689
16.567
16.622
15.278
16.431
16.978
16.786
16.712
16.611
17.211
16.695
16.557
16.799
16.627
16.675
16.802
16.844
16.726
16.181
16.692
16.495
16.480
16.647
0.074
0.278 0.281 0.047 91.234 91.201 0.528
0.096
0.272 0.281 0.064 92.343 92.338 0.615
0.146
0.288 0.292 0.104 90.724 91.487 8.011
0.309
0.275 0.275 0.045 92.936 92.288 1.485
0.083
0.253 0.252 0.066 91.998 91.966 0.425
0.057
0.351 0.357 0.043 88.181 88.290 0.417
0.048
0.291 0.294 0.039 90.399 90.450 0.461
0.160
0.327 0.342 0.070 89.975 89.918 0.804
0.118
0.270 0.268 0.062 90.952 90.904 0.723
0.112
0.310 0.316 0.050 89.359 89.122 0.661
0.134
0.289 0.294 0.074 90.190 90.250 0.465
0.105
0.286 0.286 0.056 89.489 89.463 0.457
0.540
0.336 0.303 0.134 91.755 89.530 9.98
0.093
0.296 0.296 0.061 90.043 90.042 0.297
0.101
0.264 0.272 0.074 90.881 90.746 0.546
0.142
0.268 0.267 0.070 89.449 89.547 0.713
0.108
0.305 0.282 0.143 89.139 89.013 0.604
0.114
0.276 0.280 0.058 90.261 90.159 0.743
0.057
0.282 0.283 0.043 88.054 88.023 0.326
0.097
0.296 0.289 0.063 87.688 87.692 0.471
0.145
0.275 0.283 0.058 88.905 88.819 0.78
0.194
0.311 0.318 0.065 88.930 88.983 1.04
0.054
0.291 0.274 0.069 89.962 89.923 0.426
0.097
0.253 0.244 0.045 90.476 90.290 0.652
0.361
0.226 0.261 0.045 90.721 90.337 1.473
1.419 #10.661 #10.892 1.483 117.905 117.765 7.931
Rs
Median StdDev Mean
15.692
16.575
16.655
15.279
16.438
17.015
16.824
16.726
16.664
17.232
16.651
16.588
16.793
16.633
16.722
16.798
16.857
16.758
16.149
16.731
16.513
16.487
16.661
16.605
16.791
16.708
16.837
16.770
16.117
16.568
16.453
16.672
16.446
16.677
16.575
16.278
16.786
16.919
16.748
16.636
16.843
16.467
16.425
16.705
16.730
16.925
17.035
17.052
53.417
Vm
Vmp_Median
StdDev Mean
0.035
0.021
0.085
0.022
0.066
0.115
0.112
0.087
0.118
0.073
0.113
0.097
0.081
0.074
0.121
0.024
0.093
0.108
0.093
0.108
0.048
0.029
0.065
16.628
16.827
16.747
16.965
16.794
16.123
16.562
16.496
16.680
16.494
16.738
16.606
16.257
16.751
16.966
16.673
16.658
16.889
16.471
16.413
16.679
16.724
16.939
17.041
17.197
53.465
Voc
Median StdDev Mean
20.968
21.382
21.262
21.281
21.231
21.378
21.339
21.384
21.255
21.356
21.296
21.126
21.161
21.360
21.325
21.071
21.266
21.204
21.153
21.216
21.156
21.194
21.328
0.054
0.092
0.150
0.293
0.079
0.053
0.033
0.157
0.114
0.103
0.135
0.096
0.213
0.085
0.096
0.137
0.106
0.116
0.054
0.092
0.142
0.190
0.055
0.096
0.334
0.199
Im
Median StdDev Mean
20.975
21.381
21.292
21.282
21.241
21.414
21.375
21.399
21.305
21.381
21.330
21.156
21.159
21.372
21.370
21.071
21.281
21.236
21.129
21.254
21.171
21.194
21.344
21.321
21.273
21.243
21.466
21.202
21.044
21.245
21.282
21.001
21.262
21.169
21.028
20.940
21.130
21.154
21.101
21.051
21.140
20.697
21.065
21.055
21.152
21.185
21.135
21.249
20.795
Isc
Median StdDev Mean
0.092
0.041
0.015
0.052
0.018
0.028
0.023
0.096
0.026
0.024
0.162
0.020
0.091
0.031
0.017
0.017
0.022
0.029
0.018
0.028
0.028
0.236
0.018
21.339
21.305
21.286
21.594
21.223
21.048
21.243
21.323
21.012
21.997
21.226
21.055
20.971
21.093
21.199
21.029
21.070
21.185
20.702
21.054
21.035
21.150
21.201
21.142
21.387
20.845
Module Mean
5.755
5.706
5.690
5.672
5.646
5.610
5.606
5.575
5.570
5.561
5.555
5.550
5.541
5.534
5.533
5.530
5.524
5.514
5.510
5.496
5.492
5.487
5.486
0.020
0.026
0.489
0.020
0.032
0.029
0.034
0.023
0.025
0.017
0.035
0.023
0.180
0.029
0.027
0.025
0.020
0.024
0.022
0.025
0.024
0.021
0.017
0.018
0.031
0.133
0.86
0.52
0.45
0.57
0.46
0.67
0.58
1.9
0.58
0.7
2.64
0.5
1.54
0.77
0.65
0.35
0.72
0.87
0.53
0.9
0.53
0.37
0.47
5.761
5.710
5.690
5.668
5.646
5.612
5.609
5.594
5.570
5.564
5.620
5.551
5.530
5.539
5.535
5.530
5.527
5.509
5.511
5.496
5.490
5.492
5.485
5.485
5.482
5.481
5.480
5.476
5.476
5.462
5.458
5.447
5.416
5.392
5.388
5.371
5.369
5.356
5.354
5.353
5.348
5.343
5.343
5.329
5.315
5.311
5.305
5.280
2.214
90.302
94.644
94.643
86.619
92.701
95.430
94.312
93.132
92.708
95.595
92.761
91.862
92.991
92.000
92.353
92.831
93.060
92.106
88.982
91.725
90.766
90.537
91.357
0.147
0.038
0.009
0.025
0.008
0.019
0.013
0.204
0.021
0.019
0.194
0.011
0.021
0.017
0.016
0.008
0.017
0.014
0.014
0.021
0.018
0.022
0.115
5.487
5.488
5.421
5.478
5.478
5.469
5.458
5.454
5.453
5.418
5.388
5.389
5.393
5.376
5.357
5.365
5.351
5.344
5.346
5.343
5.330
5.318
5.311
5.309
5.276
2.205
90.398
94.635
94.759
86.593
92.800
95.484
94.358
93.573
92.821
95.874
93.563
92.079
92.868
92.127
92.550
92.901
93.166
92.320
88.993
91.956
90.651
90.541
91.384
6.385
6.389
6.295
6.344
6.287
6.245
6.226
6.217
6.186
6.259
6.182
6.188
6.186
6.224
6.116
6.226
6.164
6.203
6.155
6.098
6.119
6.144
6.161
0.015
0.011
0.049
0.019
0.023
0.021
0.030
0.018
0.019
0.012
0.012
0.017
0.630
0.029
0.019
0.018
0.013
0.010
0.019
0.012
0.022
0.020
0.011
0.011
0.031
0.056
0.076
0.059
0.055
0.283
0.054
0.060
0.057
0.067
0.076
0.072
0.948
0.054
0.057
0.054
0.065
0.064
0.075
0.063
0.052
0.069
0.058
0.056
0.068
6.405
6.386
6.294
6.341
6.288
6.252
6.227
6.178
6.181
6.268
6.138
6.188
6.190
6.226
6.112
6.223
6.164
6.200
6.155
6.099
6.122
6.149
6.159
6.108
6.129
6.113
6.077
6.054
6.143
6.112
6.089
6.119
5.995
6.093
6.071
6.212
6.119
6.067
6.002
5.996
6.064
6.098
5.895
6.001
6.054
6.011
5.992
6.039
4.963
0.329
0.295
0.266
0.405
0.301
0.254
0.268
0.293
0.278
0.233
0.295
0.285
0.257
0.322
0.281
0.244
0.266
0.286
0.350
0.274
0.295
0.304
0.298
1
18
19
3
37
22
29
35
30
38
23
14
40
20
28
36
42
21
5
43
15
34
31
6.110
6.130
6.124
6.073
6.057
6.143
6.108
6.091
6.123
5.997
6.095
6.071
6.299
6.124
6.063
6.002
5.996
6.064
6.099
5.898
6.000
6.062
6.010
5.996
6.024
4.965
0.327
0.297
0.275
0.480
0.306
0.258
0.280
0.310
0.270
0.239
0.303
0.281
0.257
0.324
0.278
0.254
0.258
0.286
0.350
0.271
0.290
0.305
0.292
32
27
24
16
41
6
4
8
12
47
33
48
2
7
25
46
44
39
9
45
11
10
26
13
17
49
89.013
93.149
93.480
85.909
91.929
94.514
93.466
92.689
92.117
95.167
92.392
91.555
92.764
91.722
92.053
92.614
92.824
92.030
89.034
91.665
90.527
90.386
91.280
91.040
92.028
91.554
92.256
91.830
88.251
90.328
89.570
90.449
88.191
88.887
88.223
86.323
88.890
89.162
88.034
87.404
88.523
86.451
85.650
87.111
86.955
87.794
88.183
87.717
BROKEN!
V'=Vmp+V P'=V'*Imin
16.255
17.011
17.071
15.689
16.788
17.260
17.069
16.927
16.822
17.379
16.872
16.720
16.940
16.750
16.810
16.913
16.951
16.806
16.259
16.740
16.532
16.506
16.669
16.626
16.806
16.719
16.848
16.770
16.117
16.537
16.410
16.606
16.285
16.486
16.373
16.073
16.556
16.647
16.444
16.327
16.553
16.179
16.030
16.345
16.360
16.532
16.623
16.614
BROKEN!
FF
0.674
0.692
0.707
0.642
0.695
0.713
0.708
0.701
0.704
0.716
0.704
0.703
0.711
0.692
0.707
0.708
0.710
0.701
0.685
0.709
0.700
0.695
0.695
0.699
0.706
0.705
0.707
0.716
0.683
0.697
0.693
0.707
0.699
0.697
0.700
0.707
0.697
0.706
0.708
0.706
0.703
0.697
0.707
0.705
0.694
0.706
0.714
0.702
BROKEN!
136
Appendix F
137
Code
1
2
3
4
5
6
7
White dots
Light discoloration (brown line) around the edge of cell
Corrosion on the cell
Uneven lines on the cell fingers
Bent/damaged frame
Scratch on the glass
Dislamination/leakage
Module Nb
Comments
14
1, 4
16
1, 2, 4, 7
18 2 (many cells), 3, 1, 5
19
1, 4, 3, 2
20
6, 4, 1, 2
21
1, 4
22
1
23
1
24
1, 2
27
1, 4
28
1
29
1, 4, 6
30
4, 1, 2
31
3, 2, 5
32
1, 2, 4
35
1, 4
36
1, 2
37
1
38
1, 4
40
1, 2, 4, 5
41
1, 3, 4
42 1, 3 (whole corner)
43
1, 3
Severity
1
1
2
1
2
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
2
2
2
138
Severity values
1 Acceptible
2 Acceptible
3 Moderate
4 Significant
5 Extreme
Appendix G
139
140
Appendix H
141
Product
Description
RTSCGT3K8.1043
Theia 3.8HE-t
Power rating
AC (W)
3800
Price EXW
Drammen
NOK
We hope that you find this quotation according to your expectations, and look forward
to hearing your response.
Best regards
Espen Kristensen
ELTEK RENEWABLE
Eltek, Andyfaret 33, 4623 Kristiansand, Norway
MOBILE 90055248
Email espen.kristensen@eltek.com
Page 5 of 5
142
6840
Appendix I
143
144
145
146
Appendix J
147
Tekniske funksjonskrav
for lavspent tilknytning av
PV-anlegg
Vedlegg 3
til Tilknytnings- og
nettleieavtale for lavspente PV-enheter.
PENT
Godkjent av:
Gjelder fra:
Dok.nr.:
Utgave:
ROLJOS
BJATUF
2011-04-01
AEN-06:xxxx
1.0
148
Innholdsfortegnelse
1
Innledning ................................................................................................................................ 3
1.2
1.3
1.4
1.5
1.5.1
1.5.2
1.6
1.7
2.2
2.3
2.4
2.5
ydriftsvern ............................................................................................................................. 6
3.2
Isolasjonsniv .......................................................................................................................... 6
Side 2 av 6
149
Innledning
Den kraft som mates inn p Nettselskapets nett skal overholde de til enhver tid gjeldende
krav til spenning og effektflyt som flger av Avtaleforholdet, med mindre Nettselskapet stiller
strengere krav i det aktuelle Tilknytningspunktet. Det gjres oppmerksom p at kravene i
Avtaleforholdet er planleggingsgrenser som skal sikre at summen av flere enheter ikke frer
til at kraven i forskriften oveskrides. Gjeldende offentligrettslige krav til Nettselskapets
leveringskvalitet fremkommer for tiden i frste rekke i FOR-2004-11-30-1557: Forskrift om
leveringskvalitet i kraftsystemet. (Se kapittel 3 - Krav til leveringsplitelighet og
spenningskvalitet.)
1.2
Tillatt Spenningsbnd
For unng uakseptable stasjonre spenningsvariasjoner hos sluttbrukere, skal PV enheten ved drift ikke fre til avvik fra tillatt Spenningsbnd:
Tabell 1: Tillatt spenningsbnd i Tilknytningspunktet.
1.3
Spenningsniv [Un] :
230 V
207 V 253 V
400 V
360 V 440 V
Tillatte spenningssprang
1.4
(Maksimalt 3 %)
24
(Maksimalt 5 %)
24
0,23 kV
Un
35 kV
Tidsintervall:
1,0
95% av uken
0,8
100% av tiden
Side 3 av 6
150
1.5
THD [% av UN]
1 uke (langtid)
4,5 %
10 minutter (korttid)
6%
1.6
[% av Un]
5.4
4,5
3,2
2,7
1,8
1,4
1,4
1,4
0,9
Orden h:
3
9
15
21
>21
2
4
6
>6
[% av Un]
4,5
1,4
0,5
0,5
0,5
1,8
0,9
0,5
0,3
1.7
5
56
7
3-4
11
1,5 3
13
1 2,5
2
h )
6-8
Innmating av DC-strm
151
Overordnede krav
2.2
U >> 115
0,2
2.3
U > 110
60
U < 90
60
U << 85
0,2
Ved unormal frekvens i Mlepunktet skal PV-enheten automatisk frakobles i henhold til
kravene angitt i Tabell 8 nedenfor. Det gjres oppmerksom p at med frakoblingstid menes
tiden fra unormal frekvens oppstr til innmating fra PV-enheten opphrer.
Tabell 8: Krav til vernrespons ved unormal frekvens i Mlepunktet.
2.4
Frekvensomrde [Hz]
f > 50,2
0,2
f < 47
0,2
Side 5 av 6
152
For PV-enheter tillates det automatisk gjeninnkobling etter feil i nettet. Se Tabell 9 nedenfor
for krav til forsinkelse.
Tabell 9: Krav til forsinkelse for gjeninnkobling etter feil.
Feilvarighet [s]
2.5
< 3s
5s
> 3s
30s
ydriftsvern
Alle PV-enheter skal vre utformet p en slik mte at innmating ved ydrift ikke kan
forekomme. Ved ydrift skal PV-enheten frakobles innen 1s.
Invertere uten galvanisk skille (transformatorlse invertere) skal vre utstyrt med RCMU.
Hvis det detekteres feilstrm p AC eller DC side som er strre enn 30 mA skal inverteren
koble seg fra nettet momentant. Det skal vre to brytere i serie slik at det er redundans.
Det anbefales at det benyttes inverter med galvanisk skille.
3.2
Isolasjonsniv
Isolasjonsnivet skal vre strre enn 1 kOhm/V for invertere uten galvanisk skille.
Side 6 av 6
153
Appendix K
154
Dokumentasjon
Vedlegg 5
til Tilknytnings- og
nettleieavtale for lavspente PV-enheter.
PENT
Godkjent av:
Gjelder fra:
Dok.nr.:
Utgave:
ROLJOS
BJATUF
2011-04-01
AEN-06:xxxx
1.0
155
Innholdsfortegnelse
1!
Installatr .......................................................................................................................... 3!
2!
Enlinjeskjema .................................................................................................................... 3!
3!
Verninnstillinger ................................................................................................................ 4!
3.1!
3.2!
3.3!
Side 2 av 4
156
1 Installatr
Det skal gis informasjon om hvem som har installert de delene av anlegget som krever
godkjent el-installatr:
El-installatr:
Bjrn Hkon Lindemann
Even Evensen
Marius Waage
Agder El. Installasjon
2 Enlinjeskjema
Side 3 av 4
157
3 Verninnstillinger
3.1
Frekvens og spenning
Tabell for innstilte verdier skal fylles ut. Disse skal vre innenfor maksimale verdier bestemt
av Nettselskapet. Vernene kan vre plassert i inverteren og/eller i effektbryter.
Vernfunksjon
Maksimal verdi
Overspenning, momentan
Overspenning, forsinket
Underspenning, momentan
Underspenning, forsinket
Overfrekvens, momentan
Underfrekvens, momentan
264,5 V
253,0 V
207,0 V
195,5 V
50,2 Hz
47,0 Hz
3.2
Maksimal
frakoblingstid [s]
0,2 s
60 s
60 s
0,2 s
0,2 s
0,2 s
Innstilt
verdi
264,5 V
253,0 V
207,0 V
195,5 V
50,2 Hz
47,0 Hz
Innstilt
frakoblingstid
0,20 s
60,0 s
60,0 s
0,20 s
0,20 s
0,20 s
Beskriv type ydriftsvern (kryss av for aktivert funksjonalitet) og fyll inn innstilt verdi:
Funksjonalitet ydriftsvern
Rate of Change of Frequency (ROCOF) - df/dt
Voltage phase jump detection (Vector Shift) - d/dt
Impedance measurement
Annet (anvend merknadsfelt under for beskrivelse)
Aktivert?
Ikke kjent
Ikke kjent
Ikke kjent
Innstilling
Ikke kjent
Ikke kjent
Ikke kjent
Merknad:
3.3
Tabell for innstilte verdier skal fylles ut. Disse skal vre innenfor minimale verdier bestemt
av Nettselskapet.
Feilvarighet [s]
< 3 s (Mindre enn 3 s)
> 3 s (Mer enn 3 s)
Side 4 av 4
158
Appendix L
159
Tilknytnings- og
nettleieavtale for
lavspente PV-enheter.
(Rammeavtalen)
mellom
Agder Energi Nett AS (Nettselskapet)
p den ene siden
og
Universitetet i Agder (Innmatingskunden)
p den andre siden
(i fellesskap Partene)
PENT
Godkjent av:
Gjelder fra:
Dok.nr.:
Utgave:
ROLJOS
BJATUF
2011-02-01
AEN-06:xxxx
1.0
160
1 Partene
Partene i avtaleforholdet er:
Nettselskapet
Firmanavn
Org nr.
Postadresse
Kontaktperson
Innmatingskunden
Firmanavn
Universitetet i Agder
970546200
Postadresse
Kontaktperson
Kristen Leifsen
Stilling
Driftsleder
Tlf. kontaktperson
37233031
E-post kontaktperson
kristen.leifsen@uia.no
Dersom Partene endrer sin respektive representant skal den andre parten varsles om dette
skriftlig.
Partenes eventuelle endring av kontaktperson skal skje i trd med bestemmelsene i
vedlegg 2.
2 Avtaledokumenter
Tilknytnings- og nettleieavtalen mellom Nettselskapet og Innmatingskunden bestr av
hervrende dokument (Rammeavtalen) med flgende vedlegg (samlet betegnet som
Avtaleforholdet):
Vedlegg 1
Vedlegg 2
Vedlegg 3
Vedlegg 4
Definisjoner
Tilknytnings- og nettleievilkr for lavspente PV - enheter
Tekniske funksjonskrav
Individuelle forhold
Side 2 av 3
Rammeavtale, PV-enheter
161
Vedleggene 1-3 er i tillegg til del av Rammeavtalen, ogs Nettselskapets gjeldende vilkr for
tilknytning og nettleie for Innmatingskunder. Dersom det foreligger saklig grunn kan
Nettselskapet endre vilkrene innenfor det til enhver tid gjeldende offentligrettslige regelverk.
Nettselskapet skal informere p hensiktsmessig mte om endringer som er vesentlige for
Innmatingskunden. Innmatingskunden har plikt til etterkomme endringene, selv om disse
er omstridt. De til enhver tid gjeldende vilkr er tilgjengelig p Nettselskapets hjemmeside.
Innmatingskunden forplikter seg til holde seg oppdatert.
Ved uttak av kraft i Tilknytningspunktet gjelder Nettselskapets til enhver tid gjeldende vilkr
for tilknytning og nettleie for privatkunder p uttakstidspunktet. Nr det mates inn kraft i nettet
gjelder Avtaleforholdet p innmatingstidspunktet. I tilfellet av motstrid gis Avtaleforholdets
bestemmelser forrang.
Prioritetsrekkeflgen mellom Avtaleforholdets ulike deler fremgr av vedlegg 2 - Tilknytnings
og nettleievilkr for innmatingskunder i distribusjonsnettet.
3,8 kW
2014-04
Sted/dato:
Sted/dato:
Kristen Leifsen
Universitetet i Agder
Rammeavtale, PV-enheter
Side 3 av 3
162
Appendix M
163
164
Appendix N
Performance of Grid-connected PV
from PVGIS
165
European Commission
Performance of Grid-connected PV
PVGIS estimates of solar electricity generation
Location: 5820'3" North, 834'36" East, Elevation: 47 m a.s.l.,
Solar radiation database used: PVGIS-classic
Nominal power of the PV system: 2.1 kW (crystalline silicon)
Estimated losses due to temperature and low irradiance: 7.4% (using local ambient temperature)
Estimated loss due to angular reflectance effects: 3.1%
Other losses (cables, inverter etc.): 14.0%
Combined PV system losses: 22.8%
Ed
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Year
Total for
year
Em
1.19
2.70
4.39
6.54
7.82
8.08
7.66
6.40
4.97
2.97
1.50
0.87
4.60
Hd
36.9
75.5
136
196
242
243
238
198
149
92.2
45.0
27.1
140
1680
Hm
0.67
1.53
2.57
3.97
4.90
5.18
4.95
4.11
3.08
1.77
0.86
0.49
2.85
20.8
42.9
79.7
119
152
155
154
127
92.4
54.7
25.7
15.2
86.6
1040
Ed: Average daily electricity production from the given system (kWh)
Em: Average monthly electricity production from the given system (kWh)
Hd: Average daily sum of global irradiation per square meter received by the modules of the given system (kWh/m2)
Hm: Average sum of global irradiation per square meter received by the modules of the given system (kWh/m2)
The European Commission maintains this website to enhance public access to information about its initiatives and European Union policies in general. However the
Commission accepts no responsibility or liability whatsoever with regard to the information on this site.
This information is:
- of a general nature only and is not intended to address the specific circumstances of any particular individual or entity;
- not necessarily comprehensive, complete, accurate or up to date;
- not professional or legal advice (if you need specific advice, you should always consult a suitably qualified professional).
Some data or information on this site may have been created or structured in files or formats that are not error-free and we cannot guarantee that our service will not be
interrupted or otherwise affected by such problems. The Commission accepts no responsibility with regard to such problems incurred as a result of using this site or any
linked external sites.
Page 1/1
166
Appendix O
167
168
169