Introduction of Renewable Energy Systems in Remote

Communities in the Nordic Region

–
A Case Study of Nólsoy, the Faroe Islands

Kristian Strømmen

June 2006

Master Thesis
NTNU, Norwegian university of science and technology
Faculty of information technology, mathematics and electrical engineering
Department of electrical engineering --- NTNU 2006
Preface
I would like to take this opportunity to thank some of the people that have helped me through
this thesis.

Øystein Ulleberg, for all his help the last year during my project and master thesis, I am really
thankful for his support and guidance from day one.
Magnus Korpås, for always being supportive and positive, and for taking time to help me on
short notices.
Per Finden, for being understanding and flexible with the project shaping.
Andreas Rinnan for taking time to help me with the simulation details.
Bjarti Thomsen and Dávur Juul Magnussen for the warm welcome and helpfulness during my
trip to the Faroe Islands.
I would also like to thank NTNU, IFE and the Vestnorden Project group for their economic
support enabling me to see the project location first-hand.

i
Abstract
This thesis investigates the setup of a stand-alone energy system where the excess wind power
is diverted to and consumed in distributed domestic hot water tanks. The purpose of this setup
is to utilize wind power that would otherwise go to waste, and using it for heating tap water
and in space heating, substituting fuel oil combustion.
The first part of the thesis is devoted to local energy planning. Yearly consumption data for
electricity and oil are collected from questionnaires, statistics and calculations. Normalized
load profiles on a yearly and 24 hours basis are generated using temperature profiles and load
plots.
The second part involves building and configuring a TRNSYS simulation model for the
energy system of Nólsoy. Energy consumption data, load profiles and a local wind series are
loaded into the simulation model. Three different setups are investigated:
Scenario I is a classic wind-diesel configuration, and will be used as a reference system. With
a 300 kW windmill the diesel consumption is halved compared to a pure diesel generator
setup. However, more than 50% of the generated wind power has to be dumped.
The main idea behind scenario II is to utilize the excess wind power that would otherwise be
dumped in distributed domestic hot water tanks, one in each household (102 in total). A
number of simulations are run, now with a windmill size of 800 kW, in order to configure the
system and the tank parameters. A relatively large storage tank of 1000 litres and a moderate
heating element of 1000 W give the best energy yield; covering 75% of the tap water load and
reducing the oil consumption with over 31000 litres, 304 litres per household.
In Scenario III the tap water load is replaced with a space heating load that is more than 7
times bigger. In contrary to scenario II, a reduction of the tank size and an increasing of the
rated power yield the best results. This is a result of the large power demand, meaning that all
added power will be continuously consumed by the load, meaning a large thermal storage is
not needed. With a 50 litre tank and a 1500 W heating element, the excess wind power covers
25% of the heating load, reducing the fuel oil consumption by nearly 78000 litres, 765 litres
per household.
An economical analysis shows that scenario III is the most profitable, giving an energy price
of 0.03 €/kWh compared to scenario II’s 0.075 €/kWh. The reason is that both scenarios
require mostly the same hardware, but scenario III has greater energy utilization. Both
scenarios’ energy cost is competitive to the current fuel oil price of approximately 0.079
€/kWh.

ii
 Table of Contents
PREFACE...............................................................................................................................................................I
ABSTRACT.......................................................................................................................................................... II
TABLE OF CONTENTS....................................................................................................................................III
ACRONYMS .........................................................................................................................................................V
1 INTRODUCTION ....................................................................................................................................... 1
1.1 BACKGROUND ...................................................................................................................................... 1
1.2 THE NÓLSOY CASE STUDY .................................................................................................................... 2
2 SYSTEM STUDIES..................................................................................................................................... 3
2.1 ENERGY SYSTEM ................................................................................................................................... 3
2.2 ENERGY CONSUMPTION ........................................................................................................................ 5
2.2.1 Overview ......................................................................................................................................... 5
2.2.2 Survey.............................................................................................................................................. 5
2.2.3 Yearly energy consumption ............................................................................................................. 8
2.2.4 Tap water ...................................................................................................................................... 11
3 SYSTEM DESCRIPTION ........................................................................................................................ 13
3.1 DISTRIBUTED LOADS........................................................................................................................... 13
3.1.1 Short to medium-time storage ....................................................................................................... 13
3.1.2 Long-time storage ......................................................................................................................... 15
3.1.3 Storage system design ................................................................................................................... 16
3.1.4 Control and communication technology ....................................................................................... 17
3.1.5 Powerline Carrier ......................................................................................................................... 17
3.1.6 Wireless radio ............................................................................................................................... 17
3.2 INPUT DATA ........................................................................................................................................ 18
3.2.1 Wind power ................................................................................................................................... 18
3.2.2 Load profiles ................................................................................................................................. 19
3.3 SIMULATION SOFTWARE ..................................................................................................................... 21
3.3.1 TRNSYS ......................................................................................................................................... 21
3.3.2 HYDROGEMS............................................................................................................................... 21
3.4 TYPES ................................................................................................................................................. 21
3.4.1 Reference system ........................................................................................................................... 21
3.4.2 Wind-Diesel................................................................................................................................... 23
3.4.3 Wind-Diesel-Domestic Hot Water Tank (DHT) ............................................................................ 23
4 SYSTEM DESIGN AND BEHAVIOUR ................................................................................................. 24
4.1 ESTABLISHING A BASE CASE ............................................................................................................... 24
4.1.1 Energy source................................................................................................................................ 24
4.1.2 Loads............................................................................................................................................. 24
4.1.3 Creating the model........................................................................................................................ 25
4.1.4 Configuring the model................................................................................................................... 26
4.1.5 Simulation results.......................................................................................................................... 28
4.2 SCENARIO I: WIND-DIESEL ................................................................................................................. 28
4.3 SCENARIO II: WIND-DIESEL-DHT: TAP WATER.................................................................................. 30
4.3.1 Simulation results.......................................................................................................................... 36
4.4 SCENARIO III: WIND-DIESEL-DHT: SPACE HEATING.......................................................................... 43
4.4.1 Simulation results.......................................................................................................................... 44
4.5 LOAD CONTROL STRATEGY ................................................................................................................. 49
4.5.1 Priority level control ..................................................................................................................... 49
4.6 SYSTEM PERFORMANCE ...................................................................................................................... 53
4.7 COST OF ENERGY ................................................................................................................................ 53
4.7.1 Energy savings .............................................................................................................................. 54
4.7.2 Energy cost.................................................................................................................................... 54

iii
 4.7.3 Investment cost .............................................................................................................................. 54
4.7.4 Results ........................................................................................................................................... 55
5 CONCLUSIONS AND RECOMMENDATIONS................................................................................... 56
6 BIBLIOGRAPHY...................................................................................................................................... 58
7 APPENDIX ................................................................................................................................................ 60
7.1 APPENDIX I - SIMULATION CONFIGURATION ....................................................................................... 60
7.1.1 System settings .............................................................................................................................. 60
7.2 APPENDIX II – QUESTIONNAIRE .......................................................................................................... 63

iv
Acronyms
COE Cost of Energy
DEGS Diesel Engine Generator System
DHT Domestic Hot water Tank
ETS Electric Thermal Storage
O&M Operation and Maintenance
SAPS Stand-Alone Power System
TWS Thermal Water Storage
WECS Wind Energy Conversion Systems

v
1 Introduction
1.1 Background
The West Nordic Islands, which includes Greenland, Iceland and the Faroe Islands, are
covering a great geographical area with numerous towns and rural districts that are not
connected to the central electricity grid. These communities have their own electricity and
heat production with a local transmission network. In addition there are several sites that have
a weak and vulnerable connection to the central grid, often with a local electricity generating
unit as a supplement and backup.
The electricity and heat generation in these settlements are to a large extent based on fossil
fuel like fuel oil and diesel. A great disadvantage with these kinds of energy systems is the
society’s dependency on fuel import, which is both bothersome and costly due to the great
distances. Coupled with a steadily rising fuel price, the energy situation is becoming a
growing concern and economic load on the local communities.
Environmental issues are another important motivator to reduce the consumption of fossil
fuels, which generates emission of CO2 and other greenhouse gasses when combusted. There
will also be a risk of local oil spills when such energy carriers are used.
An increasing political focus on reliability of energy supply and environmental aspects, as
well as a growing activity on energy research and eco-tourism, has given an intensive to speed
up the work on renewable energy sources in these areas. A first step would be to utilize local
renewable energy sources, such as wind, sun, hydro- and wave power in combination with
traditional concepts like diesel and fuel oil. Future concepts with hydrogen as an energy
carrier might also be an interesting alternative.
Within this setting the Nordic Council of Ministers decided to fund research on the off-grid
energy supply in the West Nordic region, and the “Vestnorden Project” was launched at the
turn of the year 2003/2004. The project work was performed by ”Nordisk Energiforskning”,
”Institutt for Energiforskning”, ”ECON Analyse”, ”NIRAS A/S”, in addition to various local
participants on Iceland, Greenland and the Faroe Islands.
The project is separated into two parts. The first part looked at the energy situation for a
selection of localizations in the West Nordic region; some of which were investigated closer.
Local and national parties such as public officials and energy companies were included in the
process to ensure a local connection.
The first part of the project was finished in 2005, and one of the localizations chosen for
further study was Nólsoy, a small island in the Faroe Islands.
Part two of this project is currently in progress, and the main focus will be to perform a
technical and economical analysis of different system configurations based on renewable
energy.
This thesis is written with the guidance of members of the Vestnorden Project group and is
also based on results from part one of the project, but is not officially affiliated.

1
1.2 The Nólsoy case study
Renewable energy systems have a well-established history of supplying electric power in
remote, off grid areas. Such off-grid locations play an important role in technology
development as the high energy cost means that a hybrid system can become competitive
much earlier than in an urban environment. Alaska has more than 200 rural villages that have
no link to the central power grid, and are in most cases relying on electricity generated in
diesel-driven generators (Cotrell & Pratt, 2003). Because of the remoteness of these locations
the fuel and O&M prices are high, resulting in electric generation costs that average nearly
0.30 €/kWh and can be as high as 0.80 €/kWh (Drouilhet & Shirazi, 2002).
Wind-diesel systems have been utilized with success in many stand-alone energy systems for
the last decades. A fundamental problem for such systems is the renewable energy penetration
rate, since there will frequently be a mismatch between the fluctuating wind energy
generation and the load. Adding an energy storage element will in most cases improve system
performance, but large-scale electric energy storage is generally considered to be
uneconomical for such applications (Johnson el al., 2002). In communities with a large share
of fossil fuel consumption and a large heating demand, like the west-Nordic region, it might
be more economical to convert the excess power to thermal energy that can be stored
relatively cheap. This is not a new idea, and several concepts have been proposed and
implemented (Hunter & Elliot, 1994; Drouilhet, 1999; Johnson et al., 2002). Most of these
concepts use one or more big dump loads for the thermal system. A disadvantage with this
system is that depending on the location, only parts of the heating demand can be covered. It
is also limited how well one dump load can be adjusted to fit the available power.
In order to investigate the performance and concept of stand-alone systems with a thermal
dump load, the purpose of this thesis is to design and model a renewable energy system with
numerous distributed energy storage units, DHTs, for the island of Nólsoy.
The first part of the thesis is devoted to an energy analysis of the local community, where
energy consumption and load profile is generated from consumption statistics, questionnaires
and climatic data.
The second part will address the development and modelling of different scenarios with a
thermal dump load. These scenarios will be simulated in a transient simulation program and
discussed in regards to system configuration and -performance.

2
2 System studies
2.1 Energy system
This chapter is a rewrite of chapter 2 of “Decentralized energy supply based on renewable
energy sources”, 2005, by the writer.
The energy system on the Faroe Islands, like many other typical island communities, relies
heavily upon fossil fuels to cover the energy demand, because the lack of alternative energy
sources. The main fuel consumers are the fishing fleet and the stationary diesel and oil
aggregates used in electricity production. The
Faroe Islands differ however; since a substantial
part of the electricity production is based on
renewable sources, mainly hydro power.
15 %
SEV, the Faroe Islands publicly owned energy
company, runs the electricity grid and has
El
historically been the sole producer of electricity Oil
fed intro the grid. However, the Faroe Islands
Competition Authorities have ruled that the
electrical supply network is affected by the 85 %

competition laws, thus possibly allowing third

party access to the grid. This is a step on the Figure 1 - Energy cons. households, Faroe Islands
way to liberating the energy market, and can
have great impact on SEV, as well as for the Faroe island energy market.
Figure 1 and Figure 2 illustrate the Faroe
Islands’ dependency of fossil fuels, as well as
Norway’s dependency on electricity. 85%
percent of the households’ energy consumption
1%
is fuel oil, used for space heating and tap water. 15 %

With the oil price more than the double of what 5% El

Oil
it was just a decade ago (SSB, 2006), the energy
Wood
expense has become a big concern and a debated Other
topic for the Faroese. Because 60% of the
country’s electricity is generated from oil fuelled 79 %
generators, this is affecting the electricity cost as
well.
Figure 2 - Energy cons. households, Norway

Figure 3 shows an overview of the Faroe Islands energy consumption for 2001. More than a
third of the energy consumption is related to fisheries, which also contributes to 27% of the
disbursed salary for 2001. Fish and fish products is also by far the largest export article, and
amount to 98% of the country’s total export. Fluctuations in the fisheries can have a major
impact on the Faroese economy, last seen in the 1990s when a crisis in the fish industry
caused a major emigration and a serious economic decline. Since then the fish industry and
economy has recovered and is still growing, resulting in a 30% increase in energy
consumption since 1994.

3
 Road haulage
Production and 12 %
contractor Fisheries
operations 34 %
14 %

Public energy
supply Households
20 % 20 %

Figure 3 - Energy consumption per sector, the Faroe Islands (2001)

1999 2000 2001 2002 Share

Hydro 75.955 96.500 85.687 94.387 38%
Thermal 154.760 143.230 160.286 147.026 59%
Wind 504 553 2.993 7.509 3%
Total 231.219 240.283 248.966 248.922 100%
Figure 4 – Electricity production, Faroe Islands [MWh]

Table 4 shows the produced electrical energy in the period 2001-2004 by energy source
(Hagstova Føroya, 2005). The figures show a relatively high share of hydropower, as well as
a small but rising wind power production. Having a large share of hydropower is a great asset
for the energy system because it can be quickly regulated to fit the variable power output of
wind mills, without the losses of efficiency a thermal plant with the same operation would
experience. Thermal plants are still the dominant producer of electricity, covering nearly 60
percent of the demand. Most plants run on light or heavy oil, and have a power output
between 2 and 14 MW. Because such small power stations lack the efficiency and treatment
plants of bigger units, they are rather pollutive and contribute to the reason why the Faroe
Islands have one of the highest emission rates of CO2 per inhabitant in the world (UNSD,
2005). Wind power is gaining grounds, but the installed effect is still small. The total amount
of electricity produced by wind power in 2005 was 7.5 GWh, approximately 3% of the total
production. Still, this is the double of 2003 production and a sign that wind power is an area
of interest in the Faroe Islands. Wind conditions are generally very good, but there are several
disadvantages that could thwart the plans for an increase in the wind power production. The
harsh climate with strong wind gusts puts special requirements on the windmills, leading to
higher installation costs than for a normal site. Another major disadvantage for wind power is
the weak power grid, limiting the maximum installed effect. When the current projects are
finished, the total wind power installed on the Faroe Islands will be in the range of 4 MW.

4
2.2 Energy consumption
2.2.1 Overview
Gathering energy consumption data for Nólsoy proved to be demanding. In contrary to
Norway, there were no official statistics of energy consumption for regions, counties or
municipals. On the other hand, the remote and well-defined energy system of Nólsoy could
make possible a fairly accurate collection of energy consumption data.

Electricity
The electricity grid on the Faroe Islands is run by SEV, a municipality owned energy
company with a monopoly of transferring electricity. Up to recently the monopoly also
included production, but a recent change in the competition laws has opened the marked for
other companies.
Electricity consumption for households is reported by the consumers once per year, making it
easy to track the annual consumption, but a more detailed breakdown over the year is not
available. More accurate consumption data could be obtained by recording the load on the
island’s submarine cable, but this has not yet been commenced, and naturally such a study
would have to go on for some time before reliable data could be obtained.
SEV has supplied statistics of the yearly electricity consumption on Nólsoy, sorted by trade
for the most recent years.

Oil
Oil consumption data were not publicly available in the same extent as for electricity. The
main reason for this is that in contrary to electricity, the oil sale is an open market with several
competitors. For Nólsoy’s case, both Shell and Statoil had deliveries of domestic oil to private
households the last years. Naturally they were reluctant to give away sales reports. There was
also the dilemma of the privacy of the consumers to be considered.

Other energy sources

Other energy sources could typically mean solar power (thermal or electric), peat and wood
(mostly drift-wood, since the Faroe Islands have few trees). None of these sources are
monitored in any way, making it hard to estimate the amount of alternative energy sources
consumed on Nólsoy. Nevertheless, these carriers are used in a limited extent, and therefore
not included in this report.

2.2.2 Survey
Because of the limited knowledge about the energy consumption and the desire to obtain
information on the building composition on Nólsoy, it was decided that a survey could be a
good way to attain detailed user information, as well as confirming existing data. A survey
could also be a tool to include and inform the local population in the planning process,
something that was a high priority by the Vestnorden project group. Early in 2005, the author
received a survey template created by Eva Rosenberg from IFE. With the help of two local
contacts Bjarti Thomsen and Dávur Juul Magnussen, the survey was revised and adjusted for
local conditions, as well as translated to Faroese. The complete survey with results can be
found in appendix XXX. Below is a list of the key topics in the questionnaire with a brief
description.
General building information

5
 Type of house (single unit, row house), shape, size, heated area, year of construction and
type of basement, if any. First of all, these data would help categorizing the buildings into
different groups. In addition, they are needed for power demand calculations.
Heating methods
List of the household’s heating methods and its power, including floor heating. Description
of hot water boiler. These data will help map the heating infrastructure, an important
information when considering substitute energy carriers.
Thermal insulation
Thermal insulation materials and thicknesses, number and area of windows and the number
of layers in said windows. This information will mainly be used for heating power load
calculations, but will also give important information about the insulations standards, and
the possibilities for adding newer insulations as an energy-saving effort.
Electrical equipment
A list of the number typical electrical appliances, including light bulbs. An input factor in
load calculations.
Energy consumption
The total energy consumption of the most common energy sources per year and for a
typical winter/summer month. Data for other spaces, like boat houses and sheep cots have
a separate post. The electricity and oil data can be checked against the supplier’s numbers
for consistency.
A contact was made with the island’s primary and junior high school, who agreed to help with
the organisation, distribution and collection of the questionnaires. On Friday the 28th of April,
2005 the surveys were handed out to every household and business on Nólsoy (see appendix
II). A brief description of the Vestnorden project, as well as a two-page instruction was
included. It was hoped that the information regarding the Vest-Norden Project and Nólsoy’s
part in it would motivate the inhabitants to fill in the survey. The survey was anonymous, but
with a possibility for filling in name and phone number if the household agreed to be
contacted in case of any questions.
The time limit for the survey was initially intended to be 1 week. The reason for this short
deadline was the arrangement of a public information meeting late in the following week, and
it was hoped that some initial data could be presented to encourage the completion of the
survey for those who hadn’t returned it already.
The first impression from the survey was that it took longer time to receive the completed
forms than initially hoped. When the forms were picked up after scarcely a week, only 25 of
about 100 households had completed the survey. This was lower than expected. The initial
feedbacks received indicated that many people hadn’t bothered to spend time on it yet, or
were not finished. In retrospect, the design of the survey might have been one of the main
reasons for this. The form covered one full page with two columns, and probably looked more
complicated than it really was. It might have been wiser to skip some of the less important
topics, for instance “Electrical equipment”, and expanded the survey to two pages, making a
“friendlier” design. Other sections could with advantage have been compressed, for instance
the “Hot water” part, where type, installed power and age of the boiler could be written on
one line instead of covering three lines. Such changes wouldn’t degrade the information
gained by the survey, but could probably have made it more “appealing” to start working on.

6
When the survey was considered complete after one month time, 35 questionnaires had been
returned, of which 29 were for private households.
The completed surveys were added to an Access database for further study.

Survey results
Because of the limited number of businesses and public buildings on Nólsoy, this summary
will only look at the household results.
Of the 29 households that returned the survey, 3 had to be turned down because of a lack of
responses. The remaining 26 forms, corresponding to 25% of households on Nólsoy, were
fairly well filled in, but most forms had some questions that were unanswered.
Below follows a short summary of the most interesting results from the questionnaire.
• Type of building:
Single-unit dwelling: 92%
Row house with one shared wall: 4%
Row house with two shared walls: 4%
Almost all households that turned in the questionnaire were single-unit dwellings. This
doesn’t seem to quite reflect the actual building composition on Nólsoy. One reason could be
that young and middle aged people with families often live in bigger, single unit houses, while
the smaller row-houses are habited by elderly people or used as a summer house, and that the
first category is over-represented among the returned forms.
For comparison reasons, the building sizes will be sorted in four different categories
following the standard of SSB, the Norwegian Statistics Department:
• Building sizes (gross area):
Below 60 m2: 4%
60 - 99 m2: 19%
100 – 149 m2: 31%
150 m2 and above: 46%
The high percentage of buildings larger than 150 m2 is a result of the large share of single-unit
dwellings present in the survey results.
The year of construction ranged from 1650 to 2005, but 90% of the buildings were built
between 1930 and 1990. All in all the total average was 1949. The size of the heated area was
in average 72% of the total gross area. All houses in the survey had a basement, but only 11%
of those were heated.
Every house in the survey had an oil furnace, and only two households had electrical heating
utilities, including one heat pump. Unfortunately the section about thermal insulation was
lacking, but the stated insulation thickness in outer walls was spread out from 30 to 400 mm.
Because the “Windows” section suffered from the same low response rate, there was little to
comment on.
Since there will be no attempt in this thesis to estimate the household’s theoretical power
demand, the “Electrical equipment” section will not be addressed.
The average yearly electricity consumption of the households that participated in the survey
was 4053 kWh, while the average yearly oil consumption was 2964 litres, corresponding to
29933 kWh. When divided by the heated area of each household, this corresponds to 39 kWh
el per m2, and 300 kWh oil pr m2. For the net area the equivalent numbers are 29 and 219

7
kWh/m2. In comparison, Norwegian households larger than 150 m2 use in average a total of
171 kWh energy/m2 gross (SSB, 2004).

2.2.3 Yearly energy consumption

Electricity
The electricity consumption is relatively easy to monitor, since the island under normal
operation is supplied by a single, 10 kV submarine cable from the mainland. Table 1 shows
the yearly electricity on Nólsoy by sector:

Sector 2002 2003 2004 2005 2006

Fish farms 1263360 1043040 34272 42024 44460
Public buildings 165710 160726 152862 158278 156108
Building activity 7782 7848 166 1196 2072
Fishery 8 92 894 4138 3974
Transport, post and communication 98672 102098 112806 97828 96460
Street lighting 70388 61096 75956 74326 74942
Trade, accommodation and restaurant 66736 62410 63118 75996 70770
Church and bethel 2578 2430 2850 2854 2878
Agriculture 4948 4636 4328 4078 4324
Culture and spare time 6652 5846 5214 6736 6324
Boat houses 4614 3994 3018 3110 3156
Households 842844 822302 900720 859424 876392
Reconditioning etc 1444 1240 1954 1184 1194
Total 2535736 2277758 1358158 1331172 1343054
Table 1 - Total yearly electricity consumption per sector, Nólsoy [kWh] (SEV)

As the data show, the electricity consumption was almost halved in 2004 as the local fish farm
was shut down. Although there are several plans to reopen the business in some form, no
definite decisions have been made. The 2006 figures are estimated consumption.
Nólsoy has two diesel generators as a backup for unexpected interruptions in the power, for
example caused by damage to the submarine cable. The aggregates are two identical
Mercedes Benz OM404A 320 kVA 256 kW, both connected to the island’s 400V distribution
grid. One aggregate is sufficient to cover the power demand in most cases.
The backup generators are rarely used, but must be ready for operation as the submarine cable
is a vulnerable spot in the electricity network. The inlet between Nólsoy and Torshavn is
constantly being used as an anchoring ground for Russian trawlers unloading fish, and there
have been incidents where anchors have damaged the cable, resulting in power loss and costly
repairs.
One of the most important and complicated problem is to get an accurate estimation of the
maximum total power demand. For electricity, this could be achieved by measuring the power
flow on the island’s supply cable. For heating this is a more complicated issue, as there are
currently no devices installed measuring the power on oil furnaces or other fuel-based heaters.
A more theoretical approach to the maximum power load issue would be to estimate these
values using building data from the survey coupled with meteorological statistics for the
Nólsoy area. Even with a simple approximation, this method can give a good estimation that
can be adjusted against more accurate measurements of single buildings, and then applied to
larger masses of buildings. To reduce the workload it would be necessary to categorize the

8
building mass as much as possible, and accurate survey data would make this job a lot easier
and more accurate.

Conclusion
The average yearly electricity consumption per household on Nólsoy in the period 2002-05
was 8395 kWh. The total electricity consumption average was calculated from the years after
the fish farm was closed down, and corresponds to 1345 MWh.

Oil
The oil consumption on Nólsoy can roughly be divided into two sections, oil used for heating
and oil used as propellant on boats, with the latter being only a small portion of the total.
There are several available data sources for estimating the yearly oil consumption on Nólsoy.
SEV claims that a typical household on the Faroe Islands use approximately 4000 litres of oil
each year, which corresponds to about 36000 kWh of net heating if the oil is burned with 90%
efficiency. With approximately 100 habited households on the island, this would correspond
to 400000 litres of oil yearly, not including the businesses and public buildings. This is a high
estimation, compared to the fact that the average total energy consumption for Norwegian
households in 2001 was 23000 kWh (single-dwellings 27500 kWh, SSB 2004). It should be
taken into account that the Faroe Island’s climate is cool and that the number of degree days is
high, but as a comparison, Oslo has in average 4177 degree days yearly, compared to
Tórshavn’s 3600 (SSB 2005; Jacobsen). The high average wind speed could make the heating
demand higher than the number of degree days suggest. Also, the heating systems on the
Faroe Islands will have a long operating time at part load due to the cool climate, while in
Oslo the heating demand is substantially higher in the winter and almost non-existent in the
summer. This can result in lower fuel efficiency for the Faroese systems.
Because of the high electricity cost, the Faroe Island’s heating system is nearly solely based
on oil combustion. Therefore it can be assumed that in general, a household’s heating demand
equals it’s oil consumption, and vice versa. A good source for accurate numbers on the
heating demand is the district heating network outside the capital Torshavn. The system is run
by SEV, and supplies the customers with hot water supplied from a refuse incineration plant.
Table 2 shows the average heating demand per household for the district heating customers:

2000 2001 2002 2003 2004 Average

21097 20679 19863 19726 19985 20270
Table 2 - Average heating demand, district heating customers [kWh]

Calculating “backwards” with a presumed 90% efficiency in an oil boiler system, the average
heating demand corresponds to 2230 litres of oil per year. This is substantially lower than the
4000 litres SEV claims as a typical household consumption. One of the reasons for this
relatively low heating demand can be that the houses connected to the district heating network
are all new, and built with modern isolation standards. Many of these houses are row houses,
reducing the heat loss further.
Initially the two companies selling heating oil on Nólsoy were reluctant to give out sales
figures. This was partly for competitive and partly for privacy reasons. However, a couple of
files were released. The companies are referred to as company A and company B.

9
 1. The most comprehensive file was an Excel file with a register of almost 4000
different oil sales that company A delivered on Nólsoy from 1992 to 2005. Each file
is registered with an anonymous customer number, date and quantity, and was sorted
by customer. The file both includes households and businesses.
2. Company B provided a file with oil sale data from 2000 to 2004, broken up in
monthly values. The numbers were divided into 4 sectors, private, commerce, ship
and public.
3. The last document showed the total yearly sale from each company from the years
2000 to 2005.
Company A’s oil sale figures were processed in Excel, and the total yearly sales were
calculated. These values and the sum of the monthly values from company B’s file were
compared to the third file, the overview of total yearly sales. When compared, it turned out to
be a substantial error between the sum up values and the total reported sales. The reason for
this error is not known, but one possible reason is that the registration dates of sales could be
different, since Nólsoy has a central storage tank used by both companies, and this tank’s oil
level could vary from year to year.
It would be interesting to separate the households’ sale data from these numbers to compare
them with the earlier estimations based on district heating statistics from Torshavn. Company
B already had a separate post for household sales in file number 2, but company A’s sale file,
number 1, had no comments about customer type. When the total yearly sum from company
A’s file was added to the sum of company B’s household category, the yearly sum was in the
magnitude of 350000 litres yearly. Divided by approximately 100 households this would
correspond to 3500 litres yearly. This figure is a bit under the 4000 litres/year estimation from
SEV, but still a lot higher than the 2230 litres of oil equivalents consumed by the heating
network customers.
Since the company A file most likely included sales to businesses and ships, an attempt were
made to sort out the values that could not be a household purchase. Most domestic oil tanks
usually contains in the region of 1000-2000 litres, a single oil purchase larger than 2000 litres
would probably not have been to a private household. Therefore the Statoil purchase data
were re-added, but this time only oil purchases below 2000 litres were counted. This had a
considerable effect, reducing the yearly sum from between 11 to 19 percent over the 6-year
period in question. When adding these numbers to the company B household data, the 3500
litres per household per year was reduced to 3000-3200. More specifically:

2000 2001 2002 2003 2004 Average

3241 3193 3105 2969 3114 3125
Table 3 - Annual oil consumption per household, Nólsoy (litres)

The average of 3125 litres corresponds to 28125 kWh if used at 90% efficiency.
Many of the houses on Nólsoy are fairly old, build early in the 20th century, and do not follow
modern standards when it comes to insulation thickness and air tightness. This could partly
explain the considerably higher energy consumption than for instance the new houses
connected to the district heating in Torshavn.
Because of the time frame of this thesis, the simulation process had to be started before the
survey results were collected. This means that the calculated value of 3125 litres was used as

10
average oil consumption per households. Compared to the questionnaire results of 2964 litres
per year, which is 5% lower, the estimation turned out to be acceptable.

Conclusion
The oil sale statistics provided makes it possible to establish very accurately the total yearly
oil consumption. However it should be noted that the yearly consumption fluctuates
substantially, and that the different data sources don’t always give the same number. This
could be a result of big purchases being registered in two different years in different sale
statistics. Because of this, and average value of the five-year period in question would
probably be the best estimation. The yearly sales and the average are shown in Table 4:
2000 2001 2002 2003 2004 Average
410452 576358 543004 401195 423720 470946
Table 4 - Total yearly oil sales 2000-2004, Nólsoy (litres)

The average of 470946 litres of oil corresponds to 4.66 MWh (gross).

2.2.4 Tap water

The energy consumption for tap water heating is included in the oil consumption figures, but
in order to get input data for the simulation it is necessary to identify this specific energy
quantity more accurately.
Direct heating of tap water as it is used is very power demanding when the flow rate and
temperature is high. The most common solution is to use accumulation tanks for preheated
water. This reduces the power demand, and at the same time introduces the possibility of
using energy when it’s readily available or especially cheap, in example during the night time.
For an ordinary household, hot tap water will contribute to approximately 30% of the total
water consumption, corresponding to 66 litres per day per person (Hanssen et al, 1996). More
specifically:

Low value 40 – 60 l/day/person

Medium value 60 - 100 l/day/person
High value 100 - 150 l/day/person

The energy consumption for tap water heating can be expressed with the following formula:

Formula 1

Ε = q ⋅ C p ⋅ ρ ⋅ (Ttap − Tsource ) ⋅ t
where:
q: Water flow [l/s]
C p: Thermal capacity of water [4.19 kJ/kg]
ρ: Specific density of water [1000 kg/m3 ]
Tsource : Temperature of the tap water source [C]
Ttap : Desirable temperature of the tap water utilized [C]
t: Time [hours]

11
With Formula 1 the average energy consumption for tap water heating can be estimated. The
hot water temperature for ordinary households usually lies in the interval of 50-80 degrees
Celsius. The cold water is assumed to have a temperature of 5 degrees Celsius. Using these
values, heating a quantity of 66 litres per person per day would require an energy
consumption of 3.85 kWh/day per person. For a typical Nólsoy household, with an average of
2.7 persons, this would correspond to 3794 kWh/year. Other sources list average energy
consumptions ranging from 2033 kWh/year (Larsen & Nesbakken, 2005) to 3-4000 kWh/year
(Sørensen, 1977) per household for tap water alone.

12
3 System description
3.1 Distributed loads
The key duty of a wind-diesel system is to cover an instantaneous electric power load.
However, the load and available wind-power will vary over the course of the year, resulting in
shorter or longer periods when there is a mismatch between the available power and the load
demand. The wind-diesel system is designed to be able to cover the maximum load, meaning
that there will often be a surplus of power in periods of low load or high wind (or both). The
excess load has historically been disposed of in a dump load in order to maintain the energy
balance and a stable frequency and voltage on the grid. This leads to a low wind energy
penetration, as well as low fuel efficiency as a result of the diesel generators being run on
part-load.
In order to make wind-diesel systems more economical it is necessary to reduce the wind/load
mismatch. This can be achieved by storing excess energy for later use, or use loads that can be
varied according to the surplus power. Typical storage and loads for application in a wind-
diesel system can be:

Storage: Thermal storage, water pumping, hydrogen production and storage, batteries
Loads: Thermal loads, desalination, industrial load

Some of these appliances can in fact be classified both as a storage device and as a load, in
example thermal storage elements. Desalination and industrial loads were considered to be
unrealistic for the Nólsoy case study, and will not be further addressed.
It could be convenient to separate the storage part into two different chapters, short-medium
and long time storage.

3.1.1 Short to medium-time storage

Short and medium-time storage is designed to compensate for the imbalance between energy
generation and consumption on an instantaneous, hourly or daily basis. Some of these energy
storage devices are merely used to improve the quality of the delivered electricity and
stabilize the grid, allowing diesel generators, fuel cells and electrolysers to start up and power
down when needed. Such items include capacitors, flywheels and battery banks.
Additional storage devices can be added in order to improve the wind energy penetration and
reduce component sizes. In a pure wind-diesel system, some form of storage is needed in
order to maintain a reliable power supply in periods with low wind.
Another important purpose of energy storage is the possibility to even out the load on the
power grid by shifting controllable electric loads to less busy hours, for instance in the middle
of the night.

Hydrogen storage
Hydrogen storage in autonomous power grids is a relatively new but promising appliance.
Hydrogen is produced by an electrically powered electrolyser, and stored, most commonly in
compressed form. The hydrogen can then be used in a fuel cell or a hydrogen combustion
generator for electricity production. A big advantage of this setup is the ability to store large
amounts of high-quality energy over a long period of time.

13
Hydrogen storage is commonly divided into three different groups according to the
technology in question:

Pressurized gas storage

The most commonly used technology for hydrogen storage is pressurized gas storage. Until
recently these storage tanks have been fabricated from steel and aluminium with a typical
storage pressure of 200-250 bars. Research in the field of composite materials has lead to the
development of new tanks designed to withstand pressures up to 800 bars. These high
pressures are however not feasible for large-scale storage because of the rapid increase in tank
costs and the increased losses during compression.

Liquid storage
The hydrogen gas is cooled to -253 degrees Celsius, at which it turns liquid and stored in
isolated tanks at atmospheric pressure. Liquefying hydrogen is a very energy intensive
process, and energy corresponding to 28-40% of the liquid hydrogen’s heating value is lost
during the process. Another big drawback is the loss related to the vaporizing of the liquid
hydrogen under storage. This loss will vary with the size and design of the storage tank.
Typical loss values are from 0.4% per day for a 50 m3 container to 0.06%/day for a 20 000 m3
tank (Züttel).

Metal hydride storage

A chemical reaction binds the hydrogen molecules within a powder of metallic alloy,
allowing a high energy density. One advantage with metal hydride storage is the low working
pressure and low explosion and fire hazard. The biggest disadvantages are the low weight-
density and the immaturity of technology. Great progress is being made in the field of metal
hydride, and this is certainly an interesting technology for the future, especially in automotive
applications.

Conclusion hydrogen storage

Hydrogen storage is a promising technology for storing high-quality energy in large
quantities. The cost is however still too high, but is expected to fall sharply as the technology
matures. Another drawback is the low power efficiency for the total chain, typically around
20%. However, with a good system design and a wind profile that matches fairly well with
the power demand the amount of energy going through the hydrogen chain can be only a
fraction of the total energy demand over the year. A few wind-hydrogen demonstration plants
are operational, for instance the Utsira Project developed by Norsk Hydro.

Electric Thermal Storage Heaters

An Electric Thermal Storage heating unit (ETS) is a
resistance heater with an electric element encased in
ceramic blocks. When charging, the ceramic blocks are
heated up to 1200 degrees. Air is then circulated inside
the ETS, warming it to 180-200 degrees before
returning it to the room. (Skrecc, 2006) The unit will
then release the heat as needed for up to 12 hours, even
when the electricity is turned off.
Electric thermal storage heaters are usually used in
regions where the electricity is substantially cheaper
during the off-hours, and can be equipped with a

14
cooperative-installed load control device for centralized control of the unit. Units that are
coupled with an air-source heat pump are also commercially available, greatly reducing the
energy demand.
Classic room unit heating concepts typically range in sizes from 2.4 kW to 9.0 kW, with an
installed cost from approximately $1200 to $1900 (CEC, 2006).

Thermal water storage

Thermal water storage (TWS) has traditionally been most commonly used with thermal solar
collectors, but could also be useful in combination with other energy sources. In fact, a classic
domestic hot water heater can also be viewed as a thermal storage unit with a primary goal of
reducing the maximum power demand. While draining hot water the power consumption is
often in the range of 15-20 kW, while the boiler’s electric heat element could be only 3 kW or
less (Rekstad, 2000). Another advantage is the possibility of adding more power than the
current consumption, for instance from solar power or an incinerator with a fixed, optimal
power.
Water has many properties making it well-suited for thermal storage. Its specific heat capacity
is as high as 4.2 KJ/kg*K (1.16 kWh/m3*K), nearly ten times as high as iron. Another
advantage is the formation of different temperature layers in the storage tank. This ensures
that the load always will be supplied with the hottest water available, and allows for sun-
heated water to enter the tank on different levels according to its temperature.
As for ETS, thermal water storage units can be equipped with remote load control making it
an excellent appliance for controlled energy storage. One of the biggest advantages of thermal
water storage compared to ETS is its diversity. In a well designed system, a TWS can supply
both tap water and water for space heating, each at its own temperature level.

3.1.2 Long-time storage

Long time storage can, depending on the configuration, be used to store energy on a time span
ranging from a few days to seasonal variations.

Hydrogen storage
One of the advantages of hydrogen storage is the huge range in storage time span ranging
from short to almost indefinable. Long time hydrogen storage will however most likely
exclude one of the three most common storage technologies, liquid hydrogen storage. The
reason for this is the substantial storage losses over time related to evaporation of the liquid
hydrogen. These losses can be reduced a great deal for large scale storage, and has been
reported as low as 10% annual for extremely large storage vessels (thousands of m^3).
However, the size and investment costs of such tanks are not suitable for application in a
stand-alone energy system.

Long-time thermal storage

Pumped hydropower
Pumped hydro storage is a technology most commonly used by the electrical industry in
regions with a large share of thermal power plants. Thermal plants are ideally run at the rated
power, and it is an elaborate and costly job to shut them down. This means that there often
will be excess power available at night time, since the power demand is much lower than

15
during the day. This is where the pumped hydropower comes into the picture. Operating as a
normal hydropower station at the daytime, the plant has the ability to pump water back to the
storage basin at night using low-priced electricity. At daytime it runs like a classic
hydropower plant, drawing water from the basin and generating electricity for the grid.
Approximately 70% to 85% of the electrical energy used to pump the water into the elevated
reservoir can be regained. The technique is currently considered the most cost-effective means
of storing large amounts of electrical energy (Wikipedia, 2006).

3.1.3 Storage system design

For the Nólsoy case, a short to medium time storage element would be best suited to
supplement a wind-diesel system. Primarily because there are no big mismatch between the
wind and load profile throughout the year. Secondary because this case study has no ambition
of creating a pure renewable energy system. The main goal is to reduce the island’s
dependency of fossil fuel in an energy efficient and economic way.
Of the short to medium time storage systems, the two different technologies considered were
hydrogen storage and the thermal storage. After careful consideration, the choice fell on a
thermal storage system. Hydrogen systems have many possibilities in renewable energy
systems, but also some disadvantages that made them less suitable for the Nólsoy case. The
main reasons were the energy consumption profile and the already mentioned objective of the
Vestnorden project. Hydrogen technology stores high quality energy for conversion back to
electricity, but when almost 80% of the energy consumption on the island is based on
combustion of fossil fuel, the high cost and low system efficiency of a hydrogen system is not
fitting the scenario well. Instead of decreasing the diesel consumption of the electricity
generators, the gain of substituting parts of the thermal demand will be much greater from
both an environmental and a cost of energy point of view.
Of the two mentioned thermal storage technologies, the Thermal water storage is best suited
for Nólsoy, since virtually every household has a water-based heating system.

16
3.1.4 Control and communication technology
This section will give a summary of some of the most suitable communications technologies
for a distributed load system on Nólsoy.
The main goal for the load controller is to distribute the excess wind power as equally as
possible to the distributed loads. This is performed by switching the loads on and off as the
available energy varies. In order to switch the loads, a command signal has to be transferred
from the load controller to the individual distributed units. The following components are
needed to perform this operation:
1. A distributed load controller that computes and processes the control commands. The
controller must have a communications interface to both the central controller and the
distributed loads. This communication can be either one or two-way, and needs to
have the necessary transmitting rate to handle the complexity of the control signal.
2. A unit at each distributed load that can receive and intrepid the control signal sent by
the load controller, and control the load accordingly. For a two-way system the unit
must also be able to generate and transmit a return signal.
3. A communication carrier.

The most relevant communication carriers are

• Wireless radio
• Powerline Carrier
• Cable
• Fiber Optics
• Telephone line
(Johnson et al., 2002)

Wireless radio and powerline carrier were considered to be most promising technologies for
the Nólsoy system.

3.1.5 Powerline Carrier

Sending the control signal over the power grid has been used for many years in load control
applications. By connecting low-frequency radio transmitters connected to the conductors, it
is possible to transmit data with a speed up to several hundred Mbit/s (Wikipedia, 2006). An
advantage with the Nólsoy power system is that the whole island is covered by the same low-
voltage grid, meaning that the ripple signal does not have to be transmitted past any
transformers. This means that a two-way communication over the power line is possible with
at relatively low price.

3.1.6 Wireless radio

Radio transmission is a simple and robust option and that is getting more and more popular,
especially for computer networks. The technology offers two-way communication. The
specification and cost of the antenna and transmitter will depend on the range and terrain
between the load controller and the unit. Nólsoy has currently a central wireless network
transmitter that offers a high-speed internet connection, and there is a possibility that this
setup also can be used for load control communication.

17
3.2 Input data
3.2.1 Wind power
The opening of this chapter is based on chapter 4 of “Decentralized energy supply based on
renewable energy sources”, 2005, by the writer.
The Faroe Islands climate is greatly influenced by the Gulf Stream, which makes the weather
humid and windy with cool summers and mild winters. The numerous hilly islands are
causing considerable local winds, as a result of stowing, channelling and turbulence. Frequent
passing of cyclones contribute to the unstable weather conditions, with rapid pressure drops
causing damaging high wind speeds. The wind speed during a cyclone can reach as high as 40
m/s, with gusts up to 70 m/s. Average wind speeds are commonly in the range of 6-10 m/s
depending on the location. Gales are common during the autumn and winter, usually blowing
from west and south west. The wind speed is generally higher during the winter than the
summer. Though the general climate is very windy, calm periods do occur, most often in
midsummer, but then only for very short periods (Cappelen & Laursen, 2004).
There have been no long-time wind measurements on Nólsoy, but in 1989 an attempt was
made to correlate short-time wind series recorded on the island with wind series recorded in
the vicinity. Unfortunately the recordings only proceeded for two month.
In connection with part two of the Vestnorden project it was decided to initiate a long-time
wind measurement on the island. After some delay the measuring mast was raised in the
spring 2006, and the first data was collected the 10th of March. Figure 5 shows a plot of the
measured wind speed of March 2006.

Figure 5 - Wind speed March 2006, Nólsoy Wind Station (Risø/Wind-Røkt)

Due to the time span of this thesis, the recorded wind data were limited. As a compromise, a
three month series of 10 min mean values were correlated with a wind series from Mykines, a
meteorological station on the Faroe Islands. This work was done by Andreas Rinnan at IFE.
As can be seen in Figure 5, the wind speed on Nólsoy is high. The mean wind for the first
three months of measurements from the wind station was 9.39 m/s at 30 metres height, and
the yearly mean wind speed for the correlated wind series was 9.82 m/s.

18
3.2.2 Load profiles

Electricity
The method used for generating an electric load profile was similar to the procedure explained
in (Lemgart & Ulleberg, 2005):
1. Find the yearly load profile, sorted by average monthly consumption.
2. Find a representative normalized 24 hours load profile.
3. The specific load for any hour of the year can be found by multiplying the month’s
average consumption with the current hour’s normalized load.
Finding the yearly load profile for the electricity consumption proved to be challenging. The
energy company SEV could not provide any detailed graphs or statistics. In contrary to
Norway, where the electricity consumption is often reported on a quarterly basis, all private
consumers on the Faroe Islands only report once a year. The best information that could be
obtained was a claim from SEV that the electricity consumption in general was twice as high
in the winter as the summer.
In order to obtain a yearly profile it was decided that the average outdoor temperature would
serve as a basis for the curve form. Energy consumption and outdoor temperature are in many
ways connected, even though little electricity is used for heating on the Faroe Islands.
In order to obtain a higher resolution on the load profile an attempt was made to divide the
yearly profile into weekly values. The reference temperature was only available on monthly
values, so a seventh-degree polynomial function was used to estimate each weekly load ratio.
The total profile with 52 weekly load ratios was normalized to ensure an easy integration with
the simulation model. The lowest value occurred in week 31, the last week in July, where the
load was 55% of the maximum, which occurred in week 1. This fit in well with SEV’s
allegation of the summer load being half that of the winter.
The problem of limited available data reoccurred when the daily load profile was being
investigated. The only available information was a three days plot of the voltage and current
in the sub sea cable connecting Nólsoy with the main land, shown in Figure 6.

19
 Figure 6 - Voltage and current plots from Nólsoy's electricity supply cable (SEV)

The source data for the plot was not available, so the different hourly loads had to be visually
collected directly from the plot. The normalized 24 hours load curve is shown in Figure 7.

1.20
1.00
Load ratio

0.80
0.60
0.40
0.20
0.00
1 3 5 7 9 11 13 15 17 19 21 23
Time

Figure 7 - Normalized 24 hours electric load, Nólsoy

The ratio of maximum daily load divided by the minimum daily load is 2.2. This ratio and the
shape of the load curve consistent with measured load series from Grímsey, Iceland (Lemgart
& Ulleberg, 2005).

20
Heating
The load profile for the heating demand follows the same pattern as the electricity profile, as
it is highly dependant on the outdoor temperature. Since there was no additional information
available, the weekly and daily profile was assumed to be equal with the generated electricity
profile.

3.3 Simulation Software

The following chapter gives a short introduction of the simulation software and the model
types used in the simulations.

3.3.1 TRNSYS
TRNSYS is a transient systems simulation program with a modular structure, based on
FORTRAN subroutines. Originally a joint project between the University of Wisconsin -
Madison Solar Energy Lab and the University of Colorado Solar Energy Applications Lab,
TRNSYS has been developed since the early 70s. The current version, TRNSYS 16, was
released in 2004. One of the advantages of TRNSYS is its huge library of models, and the
possibility to develop new models and add them to the package.

More than 30 years after it was first released, TRNSYS is still being developed by the
University of Wisconsin, as well as several other international institutions. A library of over
400 different components has been added by developers and users.

3.3.2 HYDROGEMS
HYDROGEMS is a library of computer models for simulation of integrated hydrogen systems
based on renewable energy. HYDROGEMS is a result of more than 7 years of modelling and
simulation work performed at the Institute for Energy Technology (IFE). The models are
written in FORTRAN code, and are intended to be used with simulation programs like
TRNSYS and EES (Engineering Equation Solver), but there are also standalone versions
available. The library consists of models for power producing equipment, water electrolysis,
storage systems and control systems. Because of its generic design, system parameters
supplied from manufacturers or obtained from experiments (such as a U-I curve) can be read
to the models from an external file (HYDROGEMS, 2005).

3.4 Types
This section will give a brief description of the TRNSYS types used in the simulation, and
their key parameters. The different types are described under the system they are first
employed, but system-specific settings are listed for every scenario.

3.4.1 Reference system

The reference system for Nólsoy used in this project is somewhat different from what is
currently operating. Today, the island's electricity is delivered through a 10 kV AC sea cable
connected to the main grid at Hvitanes, close to the capital city of Torshavn. At Nólsoy, the
voltage is transformed down to 0.4 kV with a 400 kVA transformer, and fed into the local
distribution grid. Two Mercedes Benz OM404A diesel generators (DEGS) with a rated power
of 250 kW each (check this) serve as back up power should the mainland connection fail. The

21
DEGS are connected to the distribution network by 320 kVA transformers, one for each
generator.

The system configurations proposed in this paper are assumed to run as a stand-alone system,
unaffected by the central electricity grid. In order to quantify the fuel savings for these
configurations, it would be useful to compare these against an isolated power system based on
local diesel generators, rather than a grid-connected system. The base-case for Nólsoy is
therefore chosen to be a stand-alone power system based on DEGS, and the central grid
connection is neglected.

Equation block
Equations can be defined directly within the input file by using an equation component. These
equations can be used as inputs to other components, as parameters and as initial values of
inputs. An equation box can be linked to or linked from like a normal component, but rather
than being represented in the generated input file by a UNIT or TYPE statement, the
information contained will be placed in an EQUATIONS statement.

Type 9: Data Reader for Generic Data Files

This component serves the purpose of reading data at a regular time interval from a data file.
The data is converted to the desired unit and can be used by other TRNSYS-components as
time-varying force functions. Developed by the Solar Energy Laboratory, University of
Wisconsin-Madison.

Type 120: Diesel Engine Generator Set (DEGS)

This model is an empirical description of a diesel engine generator set where the fuel
consumption is expressed as a function of the normalized power output. The model can either
be supplied with a fuel consumption curve for a specific DEGS or used with the generic
model, suited for power ratings from 5-500 kW.
The generic model extrapolates from a reference fuel efficiency curve (average of 5 different
DEGS), and incorporates a correction factor derived from measurements of 20 remote area
power systems. Electrical and fuel efficiencies are calculated. The default fuel is liquid diesel,
but the model can also calculate the equivalent flow rate of 5 different fuels, including natural
gas and hydrogen.
The model is developed by Øystein Ulleberg, Institute for Energy Technology, and was
originally included in the HYDROGEMS library.

Type 102a: Control functions for Diesel Engine Generator Sets

This model contains the control function for one or several diesel engine generator sets
(DEGS) operating in decentralized power mini-grids. The DEGS are controlled in a master-
slave setting, meaning that generator i can only be turned on if generator i-1 is already on. For
each DEGS, the load power where it switches on and off can be specified. All DEGS are
assumed to be identical, and all active DEGS run at an equal power output. The model is
written by Øystein Ulleberg, IFE.

22
3.4.2 Wind-Diesel
This is a typical configuration for stand-alone systems with renewable energy penetration. On
a site with good wind condition, the fuel savings of the DEGS can be substantial. One
disadvantage of this system is that parts of the wind energy have to be dumped due to the lack
of an energy storage element.

Type 90: Wind Energy Conversion System (WECS)

The type 90 model calculates the power output of a WECS based on a power versus wind
speed characteristic loaded from an external file. The model also takes into account the impact
from air density changes and the wind speed increase with height.

3.4.3 Wind-Diesel-Domestic Hot Water Tank (DHT)

Type 4a: Stratified Storage Tank, Fixed Inlets, Uniform Losses

This type models a stratified storage tank with N (N <= 15) fully mixed layers of equal
volume. This instance of Type 4 models a stratified tank having fixed inlet positions defined
within the code. Fluid entering the hot side of the tank is added to the tank node below the
first auxiliary heater. Fluid entering the cold side of the tank enters the bottom node. The
model optionally includes two electric resistance heating elements, subject to temperature
and/or time control. The control option allows the addition of electrical energy to the tank
during selected periods of each day (e.g., off-peak hours). The auxiliary heaters employ a
temperature dead band. The heater is enabled if the temperature of the node containing the
thermostat is less than (Tset - Tdb) or if it was on for the previous interval and the thermostat
temperature is less than Tset.

23
4 System design and behaviour
4.1 Establishing a base case
The first step of the simulation process was to build a model of the basic energy system of
Nólsoy. In its simplest form, this energy system only consists of two basic components, an
energy source and a load.

4.1.1 Energy source

There are numerous energy sources and carriers utilized in the energy system of Nólsoy, the
most common being fuel oil, gasoline, gas, wood and peat. The initial studies will be limited
to the electricity consumption only. In practice, this electricity is supplied by a submarine
cable from the mainland. Since the island of Nólsoy will be considered an isolated energy
system for the bulk of the simulation, a local electricity production unit would have to be
added. Currently, two diesel generators serve as a back up system should the island’s grid
connection fail. Diesel generators will be used as base-load power source throughout the
simulation. For stand alone power systems, diesel generators have served as a foundation for
numerous of energy systems, and will continue to do so in the future. Diesel generators are
reliable and relatively cheap in acquisition, but the rise of fuel prices have resulted in a sharp
rise in operation cost. Other disadvantages with diesel generators are the discharge of green
house gases, and low efficiency on part loads and with varying operating conditions.
Since one of the motivations for the Vestnorden project was to try to reduce Nordic island
state’s dependency of fossil fuel, an alternative renewable energy source had to be considered.
Based on the unique climatic conditions in the region, wind power was chosen to be the main
source of renewable energy. Naturally it is hard to design an energy system where wind
power is the only energy source, primarily because of the fluctuations of the wind speed.
Therefore a combination of wind and diesel energy will serve as the source of electric energy
for most simulations.

4.1.2 Loads
The load represents the energy used by households, businesses and other consumers on
Nólsoy. It is important to specify the types of energy or energy carrier each load consumes,
since not all types can be substituted by each other. Thermal energy can for instance not
generally cover an electricity demand.
Since the base case energy source is purely electric, the load will initially only represent the
electricity demand on Nólsoy.

24
4.1.3 Creating the model

Figure 8 - Base case model of Nólsoy

Figure 8 shows the base case model of Nólsoy. The model has two inputs, a wind series and a
load profile, shown to the left of the picture. The wind series is a list of hourly mean wind
speeds that is representative for Nólsoy’s wind regime. These data are fed into the Type 90
WECS model, where the electricity output is calculated according to the windmill chosen in
the model configuration. Likewise, the electric load profile, consisting of hourly values of the
average electric load, is converted to the correct denomination in “BUSBAR-2” and directed
to the equation block “BUSBAR”, together with the power output from the WECS.

Control system
The busbar is the heart of the simulation model. In
short, it contains the equations that balances the
energy system and makes sure the energy demand
and energy supply match. Figure 9 shows the
control window for the component. The left
window contains the input and the right the
equations and outputs. Each time step the
following commands are performed:
1. Pload, the user load, and PWECS, the
wind power production of the time step is
collected from inputs.
2. The wind power production is compared
with the electric load. If the production is
higher, the difference is stored in the
Pexcess variable; else the shortage is Figure 9 - Control system
stored as Pdeficit.
3. If there is a shortage of power, the desired power output from the diesel generator to
obtain balance is calculated and sent to the DEGS unit (PDEGS).
4. If the power from the windmill exceeds the load, the excess power (PWECS-
Pload+PDEGSmin) is calculated and stored as the variable Pdump, the amount of

25
 excess power that must be dumped in order to keep the energy balance. PDEGSmin
is the minimum idling power output of the diesel generators in operation.

Diesel controller and mathematic model

The diesel controller takes the power demand object (PDEGS) and decides, based on the
controller settings, the appropriate control strategy for the next time step. Figure 10 shows the
settings for the controller. The
adjustable parameters are:
NMIN, NMAX: Determines the
minimum and maximum numbers
of diesel generators operating in
parallel. Note that if NMIN is set to
1, one generator will always run
regardless of the power load.
PRATED: The rated power of
each DEGS.
XLOW, XUP: Defines the “call
up” and “call down” power levels,
the load power at which the Figure 10 - Diesel Controller
corresponding DEGS is
respectively switched on and off. If the power demand is higher than the maximum power
output for the DEGS multiplied with XLOW, then a new DEGS is switched on. Likewise, if
the power demand is lower than the max power of the DEGSs multiplied with XUP, one
generator is switched off.
The controller has two outputs, PDEGS, which sets the power point for a single DEGS, and
NDEGS, the number of DEGS required to meet the load.

Simulation outputs
In order to retrieve information from the simulation, outputs from each component can be
connected to one or more output types. These types include integrators, printers, plotters and
simulation summaries. For the base case, an online graphical plotter was added in order to
visualize the power consumed or generated in the different components.
Figure 11 shows a print of the simulation results for a whole year. Note that the input data are
artificially generated, and not from Nólsoy. The red graph displays the electric load, the blue
the wind power and the pink displays the diesel generator power.
In addition to the plotter, an integrator with a file writer was connected to the most interesting
outputs. The integrator was configured to display the cumulative sum each month for each
input. The results are written to a *.out file that can be opened in most word processors,
including notepad.

4.1.4 Configuring the model

The fabricated input data from the initial simulation had to be replaced with load and wind
data generated for the Nólsoy case study.
The diesel generators currently installed on Nólsoy are not fitted for a stand-alone system.
Instead, the DEGS controller was configured for diesel generators with a rated power of 100
kW, a minimum idling power of 40kW and a maximum allowable power of 120kW. The

26
numbers of DEGS were not set, meaning that if more power is needed, another generator is
started. With a maximum power load of 265kW, this means that if the max power load occurs
at a time with zero production from the windmill, there will be three diesel generators
operating.
The effect on system performance by changing the DEGS’ rated power will not be
investigated for the base case, but it is a highly relevant parameter for a system optimization.
The windmill used for the simulation is the Bonus 300 with a rated power of 300kW. As for
the DEGS, the WECS will remain unchanged for the base case configuration.

Figure 11 - Simulation plots from TRNSYS

The load series is a combination of a hourly 24-hour load profile and a long-time average
varying with the seasons. One of the challenges is to make sure that the maximum power load
and the total yearly energy consumption (the integral of the hourly series) match. Initially,
SEV, the power company running the electricity grid claimed that the maximum power load
on Nólsoy was 110kW. However, the annual electricity consumption of 1.35GWh
corresponds to an average load of 154kW, meaning the max load must be much higher.
Seasonal variations will also affect the maximum load, and the electricity consumption on the
Faroe Islands is in average about twice as high in the winter than the summer (SEV 2006).
When compensating for seasonal variations and a fluctuating daily load profile, the
normalized maximum load would have to reach 265 kW in the coldest week of the year (first
week of January) in order to attain a yearly energy consumption of 1,35GWh. It is important
to specify that this is the normalised hourly load, and not equal to the electrical design load of
the system, since the short time load could reach substantially higher.

27
4.1.5 Simulation results
When running the base case simulations, the number of windmills in the WECS type was set
to zero. The diesel generators would then cover the complete electric load, and it would be
possible to obtain reference fuel consumption for a pure diesel system. Table 5 shows the
most important outputs from the simulation.

ELOAD EDEGS EWECS QFUEL EDUMP EDEGSdump

[MWh] [MWh] [MWh] [litre] [MWh] [MWh]
1366 1366 0 421780 0 0
Table 5 - Simulation outputs, base case

QFUEL is the total fuel consumption of the diesel generators through the whole year in litre,
and EDUMP and EDEGSdump are the amount of total surplus electricity and surplus electricity
produced by the diesel generators, respectively. Because all active diesel generators run at the
same power and reduce the power output when a new generator is started, the EDEGSdump
variable will remain zero.
The electricity consumption ELOAD is in the same magnitude as the annual consumption on
Nólsoy the last years (1358 MWh in 2004 and 1331 MWh in 2005). The diesel generators
covered the total demand without any deficits.
Dividing the total annual electricity production with the generators fuel consumption gives an
electricity generation of 3.24 KWh per litre of fuel. When using the lower heating value of
diesel, 36.5 MJ/litre, this corresponds to a fuel efficiency of 32%.
The diesel price has followed the crude oil prices to a record high level the recent years.
Currently, one litre of diesel fuel costs approximately 0.8 euro (Shell 2006). This would
correspond to a COE (cost of energy) for the base case, considering the fuel cost only, of 0.25
euro/kWh. In comparison, the electricity price to private consumers on the Faroe Islands is
currently 1.11 Danish kroner per kWh (0.15 euro/kWh)
(SEV 2006).

4.2 Scenario I: Wind-Diesel

The setup for this scenario is similar to the system End Users

illustrated in Figure 12. Electricity produced by the

windmill is utilized by the end users, supported by diesel
generators should the wind power be insufficient. In this AC

setup, one diesel generator is always running to ensure a Electricity

stable and reliable grid operation. In reality, other power
conditioning utilities like condenser banks, flywheels
and power electronics might have to be added to ensure
acceptable voltage and frequency levels. Grid stability AC

issues will not be thoroughly discussed in this thesis.

GS M
Since the WECS type was already an integrated part of
the base case model, no major adjustments had to be GS M
made in order to run the wind-diesel simulation. The Diesel Generators

number of windmills was changed from 0 to 1 in the Figure 12 - Wind-Diesel setup

WECS type, and the simulation was run. Table 6
summarizes the initial simulation for the Wind-Diesel system.

28
 ELOAD EDEGS EWECS QFUEL EDUMP EDEGSdump
[MWh] [MWh] [MWh] [litre] [MWh] [MWh]
1366 587 1545 189395 766 215
Table 6 - Simulation outputs scenario I

It should be taken into consideration that the component sizes are not optimized.
Some comments on the results:
• The wind power production was 1545 MWh, 179 MWh higher than the total electrical
load. However, only 779 MWh (50%) of this could be utilized by the consumers,
mainly because the periods where the wind production exceeded the load. This
resulted in an EDUMP of 766 MWh, corresponding to more than 56% of the total annual
consumption.
• The diesel consumption was more than halved compared to the base case; more
specifically it was reduced with 189395 litres. Because of the minimum idling power
of the diesel generators, 215 MWh of electricity produced by the diesels had to be
dumped when the wind power was high enough to cover the load demand. The amount
of dumped electricity from the DEGS corresponds to approximately 66 000 litres of
fuel, nearly 35% of the yearly consumption.
Figure 13 shows the simulation output on a monthly basis. The blue bar, ELOAD, displays the
total electricity demand each month, the red, EWECS, the total production from the windmill
and the yellow, EWECS utilized, the amount of the wind power production actually utilized by
the load.

200

150
MWh

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12
Month ELOAD EWECS EWECS utilized

Figure 13 - Simulation outputs scenario I, per month

The graph shows how the wind power production is substantially higher in the autumn and
winter months due to higher wind speeds. This matches the seasonal load variations well. An
interesting point is that EWECS utilized, the utilized amount of energy relative to the wind
power production, is remarkably stable. This is displayed in Table 7:

29
Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Des
%EWECS 56% 50% 50% 55% 51% 54% 50% 47% 48% 47% 49% 51%
Table 7 - EWECS utilization ratio, scenario I

4.3 Scenario II: Wind-Diesel-DHT: Tap water

One of the biggest challenges with a wind-diesel system is to raise the amount of utilized
wind power. The lack of storage element requires a good match between the electric load and
the wind series. Adding an energy storage element to the system can greatly increase the
system performance and wind power payload efficiency.
The motivation for introducing a hot water tank in this configuration is the possibility to
reduce the loss related to over-production of electricity in periods with high wind and low
electrical load. In a Wind-Diesel-DHT system, excess wind energy can be diverted to
distributed hot-water boilers and converted to thermal energy. As well as covering
simultaneous thermal load, the energy can be stored for later use with relatively small losses.
Although converting electricity to low-temperature heat greatly degrades the quality of
energy, this is a cheap and effective method of storing energy that otherwise would have gone
to waste.
A domestic hot water system could typically serve 1-10 households with tap water, and
possibly also supply water-based space heating.
Large loads of deviated power (on/off), such as water heaters and pumps, can give frequency
and/or voltage problems in small distributed power systems. One way of reducing this
problem, and possibly make the grid more robust, is to actively control electric loads. A
cooperative control strategy can be used on deviated loads, meaning that it is possible to turn
on hot-water boilers only when excess wind power is available, and internate between the
loads to ensure an even distribution of energy.
The wind-diesel-DHT setup is based on the wind-diesel model in scenario I. But where the
PDUMP variable represented excess energy that was dumped in scenario I, the idea is now to
utilize this energy in distributed hot water tanks (DHT). Figure 14 shows an overview of the
simulation setup.

30
 Figure 14 - Wind-Diesel-DHT setup

Since the electricity demand is already covered, the excess energy will now be despatched to
a separated thermal load controller. More specifically, the excess energy can be described as:
Pdump = max(( PWECS − Pload + PDEGS min), 0)
where PDEGS min is the minimum idling power of the diesel generators.

Thermal load controller

The thermal load controller unit is built from an “Equation Unit” type, and is the heart of the
thermal energy system. It receives an input with the available excess power PDUMP, and by
the means of the DHT rated power, the number of DHTs and the current priority system,
decides which DHTs that should power on. The outputs to the DHTs are a binary signal with
either “1” for power on, or “0” for power off.
In order to decide what DHTs to prioritize, the controller can receive a priority signal for each
DHT as an input. The n different DHTs must each have a unique priority, ranging from “n-1”
for the highest prioritized boiler to “0” for the lowest. The routines for prioritizing are solely
based on the tank temperature, meaning that the DHT with the lowest tank temperature will
receive the highest priority.
When a priority is received, the thermal load controller calculates how much excess power the
unit has available, and if this power exceeds the DHTs rated power, the unit is switched on.
Formula 2 shows the equation for calculating the available power. In short, it checks the
priority level of the DHT with an equal (eql) function, and if it matches the particular priority
in question, the PdumpUnit gets set to the corresponding power.

31
 Formula 2
PdumpUnitn =
j

∑ (eql ( DHT
i =1
n _ pri, ( j − 1)) ⋅ ( Pdump ⋅ PdumpUnitRatio))

+ (eql ( DHTn _ pri, ( j − 2)) ⋅ ( Pdump ⋅ PdumpUnitRatio − P _ DHT ))....

+ (eql ( DHTn _ pri, 0) ⋅ ( Pdump ⋅ PdumpUnitRatio − ( j − 1) ⋅ P _ DHT )

where:

DHTn_pri: The priority level for DHT n

Pdump: The total excess power available
PdumpUnitRatio: The ratio describing how much of the excess power that is available per
DHT. Default value is (1/number of households).

The control signal function is based on the

PdumpUnitn variable, which was set in
Formula 2, and is very simple:
BoilContn = GE ( PdumpUnitn , P _ DHT )
If the PdumpUnitn variable is greater or equal
(GE) to the P_DHT, the control signal is set to
1, else it is 0.
Figure 15 shows the Thermal load control
equation block. The left box shows the inputs,
the Pdump unit from the central control
system, as well as a priority signal for each of
the four connected DHTs.

Figure 15 - Thermal load control

DHT subsystem
The DHT subsystem consists of a
load, a combined storage tank and
water heater, a reheating coil and
equation blocks. Figure 16 shows a
visualisation of the cycle.

Heat load converter

A global variable called Etap_year, Figure 16 - DHT sub system
located in the “Parameter settings”

32
equation block, contains the annual energy consumption for tap water for each DHT. The
“Parameter settings” is an equation block where most variables in the simulation are
collected. This enables the user to change simulation parameters without having to click into
every different type.
The Etap_year variable is calculated by using a typical yearly consumption for a household
(3794 kWh) and multiplying it with the number of households per DHT. Initially the number
of households per DHT is set to 1. By combining Etap_year with the desired set temperature
for tap water (65 C), the average water consumption per hour is calculated using Formula 3.
Formula 3
Qavg _ hour = ( Etap _ year / 8760) / Cp _ water ⋅ (T _ tap − T _ source))
Qavg_hour: Average tap water consumption per hour.
T_source: Temperature of the water source, set to 10 degrees Celsius.
T_tap: Set point temperature for the tap water.
Cp_water: Heating capacity for water.

The daily variation in the tap water consumptions is modelled by the “Force Function” type.
A measured profile from a Norwegian housing estate is used as a basis, and the function was
normalized in order to make the daily average to 1. Figure 17 shows the profile as it is
displayed in the simulation. Note that the load profile for each DHT is shifted with 15-30
minutes from each other to avoid identical water temperatures that could cause problems in
the “Priority” controller.

Figure 17 - Tap water force function

The tap water force function is connected to the “Heat load converter” equation block. This
block has two inputs, the instantaneous value of the “Force Function”, Q_force, and the outlet
temperature of the DHT, Tload_DHT. These inputs are used in two equations. The first one,
Q_DHT, calculates the needed outlet flow from the DHT in order to satisfy the tap water load.
This function is shown in Formula 4.
Formula 4
Q _ DHTn = (Qavg _ hour ⋅ Q _ forcen ) ⋅ (T _ tap / max(Tload _ DHTn , T _ tap ))

33
In short, the equation uses the average tap water demand and multiplies this with the hourly
value to find the desired flow rate. If the tank outlet temperature is higher than the T_tap
temperature, the flow rate is reduced accordingly.
The second function, Pdemand, simply calculates the hourly power flow and makes this
available for plotting and integration.
“Preplot” is merely an equation block used to get the outputs ready for plotting and recording.

Storage tank and heater

The DHT type is a storage tank model with fixed inlets and an internal heating element.
Because the object is designed for solar heating system it has two inlets and two outlets,
allowing a separate reheating circuit, but only one inlet and one outlet will be used in this
simulation. The tank itself is stratified, consisting of n (n < 15) fully mixed layers (nodes).
The tank’s properties can be modified by changing a number of settings. The most important
are:
Parameters
V_DHT The tank’s storage capacity
Tset_DHT The set temperature
P_DHT The rated power of the heating element
DB_DHT Dead band of the thermostat
Nnode_DHT Number of nodes (temperature levels) to be used

In addition, the model has the following inputs and outputs:

Inputs
Cold-side temperature Temperature of the liquid flowing into the tank
Cold-side flow rate Flow rate of the cold-side stream
Environment temperature Temperature of the environment in which the tank is located
Control signal for element n Control signal for the auxiliary heater

Outputs
Temperature to load Temperature of the liquid flowing out of the tank
Flow rate to load Flow rate of the load stream
Thermal losses Rate of thermal energy loss to the environment
Auxiliary heating rate Average rate of power flow to the tank by auxiliary heater
Average tank temperature Average temperature of the liquid in the storage tank
Energy rate to load Rate of energy removed from the tank to supply the load

During the simulation, the tank has two inputs that change over time. Those are:
Cold-side flow rate: This input is the Q_DHT variable calculated in the equation block “Heat
load converter”. Since the tank always has a finite amount of water, the cold side flow rate
forces a similar flow rate to the load from the output Flow rate to load.
Control signal: This binary control signal is received from the “Thermal load controller”, and
decides if the tank will be on or off. Note that the tank will not switch on regardless of the
control signal if the water temperature is higher than the set point temperature Tset_DHT,
defined in “Parameter settings”.

34
Reheating coil
Since the DHT tank temperature varies with the load demand and the available excess wind
energy, the system needs a reheating coil to ensure the tap water temperature is high enough
when utilized. This secondary heating element should hold the properties of a typical peak
load heater; a high power output and a low installation cost. The operation cost is not as
important, since the DHT system should ideally cover the bulk of the energy supplied.
For Nólsoy, the choice stands between either an electric or an oil fuelled re-heater. Electric
heaters are cheap and desire little maintenance, but for this setup an oil fuelled heater was
found to be most suitable. Firstly, all households on Nólsoy already have an oil tank, so the
installation costs would be reduced compared to a full installation. Second, since the DHT is
solely dependant on electricity, it would be wise to include a secondary energy carrier to
ensure a security of supply should the electricity grid fail. This will also lower the maximum
power demand for the stand-alone system, reducing the system cost.
Inputs for the reheating coil are the mass flow and temperature of the water leaving the DHT.
The only output was the required heating rate, or more accurately the power needed to heat
the fluid to the set point temperature.

Configuring the model

The wind and electric load series were not changed from the base case model, but because of
the additional energy demand, the windmill was upgraded to an Enercon E-48 with a rated
output of 800kW.
Additional plotters and integrators were added to the most interesting outputs from the
electrical and thermal system.
The number of DHTs to be modelled was one of the key issues that had to be addressed. The
ideal solution would be to create an equal amount of DHT subsystems to match the number of
households on Nólsoy. However, this is not practically achievable with the TRNSYS
software, since all connections and outputs have to be manually connected and configured. It
was decided that a maximum of four individual DHTs were to be connected to the system at
the same time. This would make it possible to monitor each subsystem individually, and at the
same time investigate the system response to different operating and priority strategies.
Because the simulation only models a limited number of DHTs, the amount of excess wind
power available to the thermal controller was scaled down to
Pdump*PdumpUnitRatio*nDHT, where PdumpUnitRatio equals (1/number of households)
and nDHT the number of DHTs modelled in the simulation. After the simulation was run, the
outputs were scaled up accordingly so they represent the total energy system, and not just 4
units.
Figure 18 shows the complete Wind-Diesel-DHT model with all outputs connected. The solid
black lines represent actual power flow.

35
 Figure 18 - TRNSYS model, scenario II

The complete list of system settings can be found in appendix I.

4.3.1 Simulation results

Reference system
As for scenario I, the simulation was run once with the DHT units turned off and all tap water
demand served by the reheating coil. The environmental temperature T_envi was set equal to
the temperature of the water source, T_source at 10 Celsius. This temperature is assumed to
be the temperature of the room of which the tank is located, often a cold cellar on Nólsoy.
The initial temperature of the water in the DHT, T_init, was set equal T_envi. This was done
to prevent that an initial storage of energy in the DHT would be delivered to the consumers.
The simulation results are shown in Table 8.

36
 ELOAD EWECS EDEGS EDEMTAP EDELTAP EAUXTAP ELOSSTANK EREHEAT QFUEL
[MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [litre]
1366 4218 489 386 0 0 0 386 161180
Table 8 - Simulation outputs scenario II - Reference system

Below is a brief description of the different variables in the table. All energy figures are in
MWh.
ELOAD: Total electric load
EWECS: Electricity generated by the windmill
EDEGS: Electricity generated by the DEGS
EDEMTAP: Total tap water energy demand
EDELTAP: Energy delivered from DHT to the tap water load
EAUXTAP: Electricity consumed by the DHTs
ELOSSTANK: Thermal losses from the DHTs
EREHEAT: Energy consumed by the reheating coil
QFUEL: Diesel consumption by the DEGS [litre]

Default settings
The simulation was run for one year with hourly values and the default settings, but this time
with the DHTs enabled by setting the power of the heating elements to 2000 W for each
individual DHT, and a tank size of 1 m3. Table 9 shows an overview of the main outputs.

ELOAD EWECS EDEGS EDEMTAP EDELTAP EAUXTAP ELOSSTANK EREHEAT QFUEL

[MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [litre]
1366 4281 489 386 309 595 201 81.9 161180
Table 9 - Simulation outputs scenario II - Default settings

A large share of the tap water demand was covered by using excess wind power in the DHTs,
more exactly 80%. However, a very large share of the energy supplied to the DHT was lost as
thermal loss (34%). These results could imply that the thermal model of the tap water tank
was not properly configured. The storage tank was currently modelled with only one node,
meaning the temperature of the whole tank was uniform. Considering the size of the tank (1
m3), a stratification of temperature layers was likely to occur. An attempt was made to
investigate the impact of this variable, by running a series of simulations with different
number of nodes.

System design options

Number of temperature levels (nodes) in the DHTs
Table 10 shows an overview of the simulation results while varying the number of nodes.
Several of the fixed results, like ELOAD, EWECS and EDEGS have been removed from the output.

37
 Parameters Results
NNode EDELTAP %DELTAP EAUXTAP ELOSSTAP %LOSSTANK EREHEAT
[MWh] [MWh] [MWh] [MWh]
1 309 80% 595 201 34% 81.9
2 318 82% 484 107 22% 75.3
3 312 81% 442 83 19% 81.4
4 305 79% 415 70 17% 87.2
5 300 78% 395 62 16% 92.0
6 295 76% 381 56 15% 96.4
7 291 75% 369 52 14% 100
11 280 73% 340 43 13% 111
15 271 70% 319 38 12% 120
Table 10 - System outputs scenario II - Number of temperature nodes in the system

NNode: Number of nodes modelled

%DELTAP: Percentage of tap water demand covered by the DHTs
%LOSSTAP: Percentage of energy added to DHT lost as thermal loss

The number of nodes modelled had a considerable effect on the system performance. First of
all, the ELOSSTAP variable fell considerably, as expected. The reason for this change is most
likely the stratification of the tank temperature. Even when the relative level of energy stored
in the tank is low, the temperature at the top outlet will still be high enough to deliver tap
water, while the replacement water at the bottom holds a much lower temperature.
In total, the thermal loss from the storage tanks relative to the amount of energy added was
reduced with 65% when raising the number of nodes from 1 to 15.
The rate of energy delivered from the DHT to the load fell with the adding of nodes, resulting
in higher energy consumption in the reheating coil. This effect can be traced to the EAUXTAP
variable, the amount of electricity consumed by the DHT, which was reduced by over 46%
when going from 1 to 15 nodes. The stratification of the tank temperature causes the water in
the top of the tank, Tset_DHT, to reach the maximum allowable temperature (85 C) more
often, forcing the heating element to shut down even though excess wind power is available.
When comparing that the delivered energy to the DHT fell with 46%, but the delivered energy
from the DHT only fell with 12%, the effects of a stratified storage tank model becomes more
clear.
Figure 19 displays how the different variables change when the number of nodes is increased.
As mentioned, the storage tank model is written to handle up to 15 different nodes. The
question is if an even more detailed model would have a significant impact on the simulation
results. As seen on the graph, the increase in nodes has a much greater impact when going
from 1 to 2 nodes than going from 14 to 15. By visual inspection it would seem that the
impact from adding even more nodes would not alter the outputs drastically. Any change
would either way be insignificant compared to the error range in for example the energy
consumption inputs or the wind series.

38
 700

600

500 ETAPDEL
400 ETAPAUX
MWh

ETAPLOSS
300
ETAPOIL
200

100

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Number of nodes
Figure 19 – System performance versus number of temperature nodes in the storage tank

Storage tank volume

One of the most important variables in the energy system is the storage tank size. Not only
does it affect how the total system will perform, but the size, cost and properties of the tank
system will greatly affect the feasibility of the scenario, since the different households will
have the tank installed in their own homes. A tank of 3 m3 might have a very good
performance, but will most likely be too big to install in a normal household.
The geometrical shape of the tank affects its thermal properties, such as heat loss and
stratification. Initially, the tank was modelled as a cylinder with a relationship between the
height and the radius of two to one, meaning that the height was twice the radius independent
of the tank size. This ratio will give the lowest possible surface area, reducing thermal losses,
and the ratio will be used for the tank size sensitivity analysis.
The tank will be modelled with 15 thermal nodes for the greatest possible accuracy. Table 11
shows the most important outputs from the simulation.

Parameters Results
VDHT r/h DHT EDELTAP %DELTAP EAUXTAP ELOSSTANK EREHEAT %REHEAT
[litre] [MWh] [MWh] [MWh] [MWh]
150 0.29/0.58 179 46% 187 8 210 54%
250 0.34/0.68 199 52% 212 13 188 49%
500 0.43/0.86 245 63% 272 24 143 37%
750 0.49/0.98 258 67% 295 30 131 34%
1000 0.54/1.08 271 70% 320 38 120 31%
1250 0.59/1.17 277 72% 337 46 114 30%
1500 0.62/1.24 281 73% 352 53 110 28%
2000 0.68/1.37 287 74% 376 66 104 27%

39
 3000 0.78/1.56 295 76% 419 89 96 25%
Table 11 - System outputs scenario II - Storage tank size

VDHT: Tank size of DHT [litre]

r/h DHT: Radius and height of the DHT tank [m]
%REHEAT: Energy demand in the reheating coil as a percentage of the total tap water
energy demand

Figure 20 displays how the oil consumption is reduced with the different tank sizes compared
to a pure externally heated system. In general, the larger the tank the more oil can be saved,
but the ratio falls when the tank size increases. At one point the process will be reversed as the
thermal losses will grow too big, but it was chosen to limit the tank size to 3 m3 since bigger
sizes appeared as totally unrealistic.
As can be seen in Table 11, larger tank sizes yields more thermal loss and demands more
excess energy from the grid. The ratio of loss versus gain grows bigger with increased tank
size. For example, increasing the tank size from 1500 to 2000 litres reduces the oil
consumption with 1%, but the electricity consumption increases with 7%.
Eventually the question boils down to if the extra oil savings can economically justify a larger
tank investment, and if the extra excess power has an alternative cost, meaning if it can be
utilized in other parts of the energy system. This will be discussed later in the thesis.

80 %

70 %
Reduction of oil consumption

60 %

50 %

40 %

30 %

20 %

10 %

0%
150 250 500 750 1000 1250 1500 2000 3000
DHT size [litre]

Figure 20 - Oil consumption versus DHT tank size

Maximum heating rate of DHT

Another crucial variable in the DHT sub system is the maximum heating rate of the DHTs.
The reason for this variable’s importance is the way the excess energy is divided to the
heaters. Let’s say there is excess wind power of 10 kW available for the four DHTs. If the
rated power of each DHT is 2kW (default setting), then all four heaters would be switched on,
but 2 kW would go to waste. If the rated power on the other hand was 3kW, then three heaters
would operate, and 1kW would go to waste. In short, low rated power of the DHTs will
ensure that they are switched on often, but lots of energy would go to waste if the power
available is high. On the other hand, a large rated power would mean that a lot of energy

40
could be utilized when the excess power was high, but at low power surplus you might not be
able to turn on any of the heaters at all.
Table 12 shows the results of a simulation where the rated power of the DHTs, the P_DHT
variable, is varied.

Parameters Results
PDHT EDELTAP %DELTAP EAUXTAP ELOSSTANK EREHEAT %REHEAT
[W] [MWh] [MWh] [MWh] [MWh]
500 257 67% 305 35 132 34%
750 286 74% 343 43 107 28%
1000 290 75% 348 44 104 27%
1250 288 75% 345 44 106 27%
1500 285 74% 340 43 109 28%
2000 271 70% 320 38 120 31%
2500 252 65% 298 36 137 35%
3000 240 62% 283 34 148 38%
4000 213 55% 251 31 178 46%
5000 185 48% 218 27 204 53%
Table 12 - System outputs scenario II – Maximum heating rate of DHT

350
Energy delivered to tap

300
water load [MWh]

250
200
150
100
50
0
500 750 1000 1250 1500 2000 2500 3000 4000 5000
Unit power [W]

Figure 21 - Energy delivered to tap water versus maximum unit power of DHT

Figure 21 shows that the maximum energy utilization occurs when the heating element is
around 1000 W. However, the amount of energy utilized changes only with one percent when
the element size is varied from 750 to 1500 W, making it hard to conclude on an “optimal”
element power. Most likely other system settings will affect this ratio, such as the tank size. It
seems clear however that elements bigger than 1500 W will decrease the system performance
at the current settings.

41
Temperature settings
There are mainly two temperature settings that can be controlled and that could affect the
system performance, the tank set point temperature Tset_DHT, and the dead band temperature
DB_DHT.
The tank set point temperature describes the maximum allowable tank temperature. A high set
point temperature means that more energy can be stored in the tank, but will also increase the
thermal losses. To avoid the possibility of boiling, this temperature cannot be set higher than
95 degrees Celsius, and since the tap water demand is at 65 C, this will be the minimum limit.
The default Tset_DHT is 85 C.
The dead band temperature describes how far the tank temperature will fall below the
Tset_DHT temperature before the heating element is turned on again (given a positive control
signal), and the main reason it’s being utilized is to prevent a rapid rate of switching of the
heat element. The default dead band is 5 C.
The simulation was run with a tank size of 1000 litres and a 1000 W heating element. The
outputs can be seen in Table 13.

Parameters Results
TSET_DHT TDB_DHT EDELTAP %DELTAP EAUXTAP ELOSSTANK EREHEAT %REHEAT
[C] [MWh] [MWh] [MWh] [MWh]
95 2.5 298 77% 364 50 100 26%
5 298 77% 363 50 100 26%
7.5 297 77% 362 49 101 26%
90 5 294 76% 356 47 102 26%
85 2.5 291 75% 349 45 104 27%
5 290 75% 348 44 104 27%
7.5 289 75% 347 44 105 27%
80 5 286 74% 340 42 107 28%
75 2.5 282 73% 333 39 109 28%
5 281 73% 331 39 110 28%
7.5 279 72% 328 38 110 28%
70 5 275 71% 321 36 113 29%
65 2.5 269 70% 312 34 118 31%
5 266 70% 308 33 120 31%
7.5 260 67% 301 33 126 33%
Table 13 - Simulation outputs scenario II - DHT tank temperature and dead band

The dead band temperature was changed between three different levels, 2.5, 5 and 7.5
degrees. Because of the large number of simulations, only half of the different set
temperatures were simulated with a varying dead band.

42
The simulation results showed that the temperature level and dead band had a relatively small
impact on the amount of energy delivered to the load. With a dead band of 5 C, the change in
delivered energy was only 7% when reducing the tank temperature from 95 to 65 degrees
Celsius. Adjusting the dead band while holding the set temperature constant resulted in a
change in the delivered energy by less than two percent for all set temperatures, except 65 C.
All in all the set temperature and the dead band temperature had little effect on the amount of
energy delivered, with the highest set temperature and the lowest dead band being the most
favourable from an energy exploitation ratio point of view. With a simulation time of one
hour, the effect of rapid power switching of the element due to the dead band temperature can
not be investigated.
As a conclusion, technical issues such as the storage tank’s recommended temperature level
and dead band should be deciding when setting the tank parameters.

Summary
Varying the tank size had a noticeably effect on the system performance, and a bigger tank
gave a higher energy yield. However, the ratio of useful energy to added energy fell with the
increase in tank size as the thermal losses rose, and the investment cost will also be higher
with a bigger tank. 1000 litres was found to be a suitable size with high energy utilization but
reduced thermal losses.
Unlike the tank size, the maximum heating rate of the DHT had a defined optimal interval
from 750 to 1500 W. The performance difference within this interval was minimal, so 1000
W was chosen as a basis.
Changing the tank temperature had small effects on the system, but in general the higher
temperature will give a better performance. The set point was chosen to 85 C to avoid any
raised investment cost related to a very high tank temperature.

4.4 Scenario III: Wind-Diesel-DHT: Space heating

The introduction of a DHT subsystem to utilize excess wind power for tap water heating
showed promising simulation results. At best, over 75% of the tap water demand was covered,
corresponding to a decrease in oil consumption of almost 300 MWh. The amount of utilized
wind power with the 800 kW turbine was raised from about 900 MWh (21%) to 1200 MWh
(27%). However, the greater part of the wind energy still remains unused. Considering the
electricity and tap water load on Nólsoy only corresponds to about 27% of the total energy
consumption, there exists a great potential for utilizing more wind power as a substitute for
fossil fuel.
Most of the remaining energy consumption is oil-based space heating in the households. The
idea of scenario 3 is in many ways similar to scenario 2; using distributed domestic heating
tanks that utilize excess wind power, but this time to cover the space heating load, which is
roughly 7-8 times as large as the tap water load.
The energy consumption for space heating per household was estimated to 28125 kWh per
year. Multiplying by the number of households, this sums up to 2.9 GWh.
The simulation setup is in general equal to the tap water setup in scenario II. The excess wind
power is distributed to DHTs where it is converted into thermal energy and stored. Hot water
is drawn from the top of the tank to cover the heating demand, and is reheated in a separate
reheating coil should the temperature be below the desired temperature in the radiators,
T_radiator. The major difference from the tap water setup is that the heated water returns to
the DHT after giving off heat in the radiators. This required a redesign of the DHT subsystem.

43
DHT subsystem
Figure 22 shows a picture of the
redesigned DHT subsystem.

Heating load
The thermal controller was completely
redone. In contrary to the tap water load,
that is assumed to have a constant profile
throughout the year, the heating demand
fluctuates due to the seasonal temperature
change. A week load profile similar to the
electric load was combined with the daily
load ratio to form a complete yearly
profile of hourly values. To cover the heat
demand, the load controller used the water
temperature from the reheating coil as an Figure 22 - DHT sub system
input, and with the help of the specific
hourly power demand calculated the output temperature from the load, which was directed
back to the DHT. The water flow rate was defined as a constant value, chosen high enough to
prevent the output temperature from the thermal load to fall below room temperature; else the
heat energy could not be transferred to the load. The calculated power demand (Formula 5)
and the actual power flow (Formula 6) were directed as outputs to enable an energy balance
check.
Formula 5
Desired power demand:
Pdemand = WeekLoad ⋅ HourLoad ⋅ Pavg _ hour

Actual power delivered:

Pdelivered = Qradiator ⋅ Cpwater ⋅ (Tinradiator − Toutradiator )
Formula 6

WeekLoad: Relative load of the week

HourLoad: Relative load of the hour
Pavg_hour: Average heat load over the year
Qradiator: Water flow in the load circuit
Tinradiator: Temperature in to the load unit
Toutradiator: Temperature out of the load unit

4.4.1 Simulation results

Reference system
For the reference system, the DHT heating elements were switched off in order to get a
reference value for oil consumption. The simulation results can be seen in Table 14, all results
in MWh if something else is not specified. For a complete list of settings, see appendix I.

44
ELOAD EWECS EDEGS EDEMHEAT EDELHEAT EAUXHEAT ELOSSTANK EREHEAT QFUEL
[MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [litre]

1366 4281 489 2882 -116 0 166 3000 161180

Table 14 - Simulation outputs scenario III - Reference system

EDEMHEAT: Total heating demand, households

EDELHEAT: Energy from the DHT delivered to the heat load
EAUXHEAT: Electricity consumed by the DHTs
ELOSSTANK: Thermal losses from the DHTs
EREHEAT: Energy consumed by the reheating coil
QFUEL: Diesel consumption by the DEGS [litre]

The windmill and electricity load settings were not changed from scenario II, meaning the
ELOAD, EWECS, EDEGS and QFUEL variables remain unchanged. EDEMHEAT, the total space heating
demand, corresponds to an average consumption of 28125 KWh per households. EDELHEAT,
defined as the difference in the energy of the water flowing out of the tank from the water
flowing into the tank, is lower than the thermal losses because of the difference in the energy
storage level in the tank from the start to the end of the year.
The thermal losses from the storage tank have grown substantial, mainly because of the high
average temperature in the tank. Since the water from the radiators is returned to the tank at
the bottom node, and the flow rate is rather high compared to the tap water scenario, the
stratification effect is much less dominant resulting in increased thermal losses. It must be
taken into account that the energy flow through the system is much higher than in scenario II,
and hence the losses increase accordingly.
EREHEAT is the energy amount added in the reheating coil to satisfy the thermal load, and
equals the heating demand variable plus the difference in the tank’s energy storage level.

Default settings
The heating elements were then activated and a simulation was run with the same parameter
settings as the reference case, with a tank volume of 1000 litre and heating element of 1000
W. The results are displayed in Table 15:

ELOAD EWECS EDEGS EDEMHEAT EDELHEAT EAUXHEAT ELOSSTANK EREHEAT QFUEL

[MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [MWh] [litre]
1366 4281 489 2882 398 550 168 2484 161180
Table 15 - Simulation outputs scenario III - Default settings

A total of 398 MWh of energy was added to the DHTs and utilized in the load. This is
approximately 100 MWh more than the best tap water scenario, but with a utilization rate of
the wind energy of 30%, the potential for improvement is still great.

45
 Figure 23 - Simulation plot, scenario III

Figure 23 shows a plot from the simulation. The turquoise graph displays the temperature of
the water entering the radiator, while the pink graph displays the temperature of the top node
of the storage tank. As the graph shows, the tank temperature is never in the range of the
upper set point temperature of 85 C, so no available energy is rejected from the DHT due to
full storage capacity (which frequently happened in the tap water simulations).
In order to improve the wind energy utilization rate, a number of simulations were run while
varying key variables.

System design options

Storage tank volume
Since the energy storage level never was completely utilized during the initial simulation, a
change of the tank size could possibly reduce the thermal losses and improve the system
performance. A number of simulations were run while varying the storage tank size. The
simulations are summarized in Table 16.

Parameters Results
VDHT r/h DHT EDELHEAT %DELHEAT EAUXHEAT ELOSSTANK EREHEAT %REHEAT
[litre] [MWh] [MWh] [MWh] [MWh]
50 0.20/0.40 532 18% 550 23 2351 78%
100 0.25/0.50 522 18% 550 37 2360 79%

46
 150 0.29/0.58 514 18% 550 48 2369 79%
250 0.34/0.68 498 17% 550 68 2385 80%
500 0.43/0.86 463 16% 550 107 2420 81%
750 0.49/0.98 429 15% 550 139 2453 82%
1000 0.54/1.08 398 14% 550 168 2484 83%
1250 0.59/1.17 369 13% 550 194 2513 84%
1500 0.62/1.24 341 12% 550 219 2541 85%
2000 0.68/1.37 291 10% 550 264 2592 86%
3000 0.78/1.56 203 7% 550 343 2679 89%
Table 16 - Simulation outputs scenario III - DHT tank volume

The system was simulated with storage sizes down to 50 litres.

In contrary to the tap water scenario where the amount of utilized energy increased with an
increase in tank size, the opposite was the case for scenario III. The reason can be seen from
the EAUXHEAT variable, which remained constant throughout the simulation. This confirms the
assumption that the storage level is oversized, or seen from a different view, the excess energy
added is too low. Even with storage tank of only 50 litres the added excess wind power could
never bring the tank temperature above 65 C, far below the set temperature of 85 C.

Maximum heating rate of DHT

In scenario II, the storage tank had trouble absorbing the excess wind power in periods with
high wind, because the thermal storage capacity was full. The solution that yielded the best
energy utilization ratio was to reduce the element power in order to increase the operating
time.
For the space heating case in scenario III, the problem is reversed. All excess wind energy is
added to the heaters, but the goal is to raise the available power. This will be attempted by
increasing the rated power of the heating elements at the sacrifice of operating time.
The simulation was initialized with a storage tank of 500 litres.

Parameters Results
PDHT EDELHEAT %DELHEAT EAUXHEAT ELOSSTANK EREHEAT %REHEAT
[W] [MWh] [MWh] [MWh] [MWh]
500 217 8% 307 106 2666 89%
750 346 12% 434 107 2537 85%
1000 463 16% 550 107 2420 81%
1250 564 20% 649 107 2319 77%
1500 641 22% 724 107 2241 75%
1750 611 21% 689 107 2273 76%
2000 629 22% 703 107 2256 75%

47
 2500 543 19% 609 107 2351 78%
3000 595 21% 660 107 2317 77%
4000 425 15% 480 107 2509 84%
5000 484 17% 539 107 2487 83%
6000 482 17% 536 107 2527 84%
Table 17 - Simulation outputs scenario III - Maximum heating rate of DHT

700
Energy delivered to space

600
heating load [MWh]

500
400
300
200
100
0
500 750 1000 1250 1500 1750 2000 2500 3000 4000
Unit power [W]

Figure 24 - Energy delivered to space heating load versus DHT unit power

Figure 24 shows a plot of the energy delivered to the space heating load with different unit
power sizes. The utilization rate rose nearly linear from 500 up to 1500 W, where the peak
was reached at 641 MWh, 22% of the heating demand. From 1500 to 3000 W the power
output was fluctuating, but still at a high level. The energy variation in this interval was most
likely a result of random match between the excess wind power and the unit power, and with
another wind profile from the same site but another year, the results might have been
different.
If a choice had to be made between two different unit powers with nearly the same energy
yield, for instance 1500 and 2000 W, the smaller one would most likely be better from an
operational point of view. A smaller rated power would mean a longer operating time and
lower power spikes when switching, improving grid stability and quality.
A few additional simulations were run to study the system behaviour with a heating element
of 1500 W with varying storage tank size. As Table 18 shows, the storage capacity was not
reached and hence the lowest storage size gave the best energy yield.

48
 Parameters Results
PDHT VDHT EDELHEAT %DELHEAT EAUXHEAT ELOSSTANK EREHEAT %REHEAT
[W] [litre] [MWh] [MWh] [MWh] [MWh]
1500 50 708 25% 724 23 2176 73%
1500 100 699 24% 724 37 2184 73%
1500 150 691 24% 724 49 2192 73%
1500 250 677 23% 724 68 2207 74%
Table 18 - System outputs scenario III - DHT power and tank volume

Temperature settings
For scenario III the choice was made to not run any simulation with varying temperature
settings. The reason for this was that the thermal storage capacity was never reached, in other
words there would be no reason to increase the set point temperature since the current
boundary was not limiting.

Summary
In contrary to the tap water scenario, where the rate of utilized energy rose with increased
tank size, the opposite is the case for the space heating setup. The reason for this is the high
power demand, meaning that all energy added to the system will be consumed by the load.
Adding storage will only increase the thermal losses, and hence decrease system performance.
In conformity with scenario II, the heating element size had an optimal interval, but for
scenario II this range was from approximately 1500 to 3000 W. 1500 W was chosen as the
favourable setting, not only because it gave the highest energy output, but because a lower
element size will give a longer working time and smaller load spikes, improving grid stability.

4.5 Load control strategy

The purpose of this section is to examine how control strategies for the excess wind power
will affect the energy yield and system behaviour in general.
For all simulations performed in the previous chapters, the load priority was simple; the DHT
with the lowest tank temperature was first in line to receive any excess power, the second
lowest was number two in line, and so on. This control strategy assumes a two way
communication, no time delay and perfect temperature readings from the tank.
The biggest drawback with this control strategy is the requirement of a two-way
communication utility. This is achievable in practice for most cases, but will demand
additional hardware, increasing investment costs. In some cases, it might not even be
technically possible to utilize a two-way system. It would therefore be interesting to
investigate the performance of a one-way control system.
The tank parameters used in each scenario are as follows:
Scenario II: Tset_DHT = 85 C, P_DHT = 1000 W, V_DHT = 1 m3
Scenario III: Tset_DHT = 85 C, P_DHT = 1500 W, V_DHT = 0.05 m3

4.5.1 Priority level control

The key goal for the control system will be to spread out the excess wind power as equally as
possible between the DHTs. In order to achieve this it is necessary to develop a priority

49
system that decides what units are highest on the list to receive power when it becomes
available, and to shift this priority at a given time interval.
The priority level is shifted by the load controller, and must be transmitted with the control
signal. Each DHT must be able to interpret and identify its own, unique priority level.
Since the TRNSYS model only consists of four independent sub-systems, each of them are
assumed to consist of one fourth of the total amounts of DHTs, and share the same priority
level. This means that there will be four different priority levels.
The first rounds of simulations were run with a priority shift ever hour, every three hours and
lastly every six hours. One hour was the lowest possible setting considering the simulation
time.
Figure 25 shows the one-hour priority function for the first of the four DHTs. The average
value of each hour defines the priority, and the function is looped every four hours. DHT nr 2
starts at a level of 1 at time 0, DHT 3 starts at 2 and DHT 4 starts at 3.

Figure 25 - DHT priority force function

System results
Scenario II:
Control settings EDELTAP %DELTAP EAUXTAP ELOSSTANK EREHEAT %REHEAT
[MWh] [MWh] [MWh] [MWh]
Temperature control 290 75% 348 44 104 27%
One hour priority shift 290 75% 349 44 104 27%
Three hour priority shift 290 75% 347 44 105 27%
Six hour priority shift 289 75% 347 44 106 27%
Table 19 - System outputs load control strategy - Scenario II

50
Scenario III:
Control settings EDELHEAT %DELHEAT EAUXHEAT ELOSSTANK EREHEAT %REHEAT
[MWh] [MWh] [MWh] [MWh]
Temperature control 708 25% 724 23 2176 73%
One hour priority shift 708 25% 724 23 2176 73%
Three hour priority shift 708 25% 724 23 2176 73%
Six hour priority shift 708 25% 724 24 2177 73%
Table 20 - System outputs load control strategy - Scenario III

As Table 19 and Table 20 shows, the change in priority control had close to negible effect on
the total system performance for all three time shift settings. This was the case for both the tap
water and the heating load simulations.
Scenario II and scenario III are similar in setup, but the systems operate very differently.
Scenario II, the tap water load, has a large renewable energy penetration (~75%) and relies on
a large storage capacity in order to utilize the excess wind power. Scenario III on the other
hand, with a space heating load, has a high power demand and a relatively low share of
renewable energy (~25%). Since the energy added to the system can be utilized almost
instantly all year around, there is no need for large energy storage.
The reason scenario II responds so well to the change in priority time, in addition to the
storage capacity, is the low heating element power. When the heating element is operating
uninterrupted, it takes almost 12 hours to raise the tank temperature by 10 degrees Celsius,
even when there is no energy flow to the load. With an average tap water load drawn from the
tank at the same time, the time of heating the tank 10 C is raised to over 20 hours. Similarly it
takes 27 hours to lower the temperature by ten degrees with an average tap water load drawn
from the tank, without considering thermal losses.
For scenario III, the heating demand alone is enough to absorb the energy added by the DHTs
heating element. The average heating demand throughout the year is 3.2 kW, and in the
warmest summer month the daily average is 1.7 kW. Compared to the maximum power of the
heating element, which is set to 1.5 kW, this means that the excess power will never “go to
waste” no matter the priority control, as long as the correct amount of DHTs are given a
positive control signal.

DHT behaviour
All the simulations above were run with a nearly identical load profile in all DHTs. It would
be interesting to examine how a difference in the load demand would affect the system
behaviour, but most importantly the individual DHTs.
Because of the robust system response of the space heating system, these simulations will
only be performed for scenario II, the tap water load, and only with 3 and 6 hour priority
shifts.
To adjust the load demand, the “Force Function” type in each DHT was adjusted individually.
The load profile of the first and the last DHT was lowered and raised with 20% respectively,
while DHT 2 and 3 were adjusted by 10% in an identical fashion. In order to monitor each
DHT more closely, the energy demand and energy delivery of each unit were connected to an
integrator unit.

51
 EDEMTAP %DEMTAP EDELTAP %DELTAP EAUXHEAT EREHEAT
[MWh] [MWh] [MWh] [MWh]
DHT1 76.4 19.8% 58.5 20.2% 73.1 19.5
DHT2 86.9 22.5% 65.5 22.6% 80.1 23.4
DHT3 107.0 27.7% 79.9 27.6% 94.3 29.2
DHT4 116.4 30.1% 85.9 29.6% 100.2 32.7
Total 386.7 100.1% 289.8 74.9% 347.7 104.8
Table 21 - System outputs differated loads - Temperature priority

EDEMTAP %DEMTAP EDELTAP %DELTAP EAUXHEAT EREHEAT

[MWh] [MWh] [MWh] [MWh]
DHT1 76.4 19.8% 59.2 20.5% 74.2 19.0
DHT2 86.9 22.5% 66.5 23.1% 81.4 22.5
DHT3 107.0 27.7% 78.7 27.3% 92.5 30.5
DHT4 116.4 30.1% 83.9 29.1% 97.5 34.7
Total 386.7 100.1% 288.3 100% 345.6 106.7
Table 22 - System outputs differated loads – 3 hour priority shift

EDEMTAP %DEMTAP EDELTAP %DELTAP EAUXHEAT EREHEAT

[MWh] [MWh] [MWh] [MWh]
DHT1 76.4 19.8% 59.9 20.8% 75.4 18.3
DHT2 86.9 22.5% 67.1 23.4% 81.4 21.9
DHT3 107.0 27.7% 77.3 26.9% 91.1 31.9
DHT4 116.4 30.1% 83.0 28.9% 97.0 35.6
Total 386.7 100.1% 287.3 100% 344.9 107.7
Table 23 - System outputs differated loads – 6 hour priority shift

As Table 21, Table 22 and Table 23 shows, the effect of adjusting the load balance had an
insignificant impact on the system. Even with a 6 hour priority shift, the DHTs had almost the
same EDELTAP, the amount of useful energy delivered, as for the temperature controlled case.
The main reasons for these results are most likely that the tank storage capacity had been
limiting the energy absorption at these settings. When the unit’s consumption was increased,
so was its ability to receive power. If the storage size had not been limiting, then all DHTs
would in theory receive an equal amount of energy (depending on the wind), and the ones
with the highest energy consumption would have had a poor utilization rate of renewable
energy.

52
4.6 System Performance
A summary of all the three simulation scenarios is listed in table. For scenario II and III, the
parameters have been chosen to return the highest possible energy yield. One exception from
this is the Tset_DHT temperature of scenario that was set to 85 C instead of 95 C. The reason
for this is the small difference in energy output (2.7%), and the uncertainty about all DHTs
ability to run continuously at such high set temperature. For energy conversion between kWh
and litres of diesel, an efficiency of 90% is used.
Scenario I: PWECS = 300 kW
Scenario II: PWECS = 800 kW, Tset_DHT = 85 C, P_DHT = 1000 W, V_DHT = 1 m3
Scenario III: PWECS = 800 kW, Tset_DHT = 85 C, P_DHT = 1500 W, V_DHT = 0.05 m3

Priority Average Electric load Diesel Heating/tap water Useful wind

System control penetration served by savings inc. load served by energy [% of
[%] wind [%] thermal [%] wind [%] total wind]

Scenario - 113 57 55 - 50
I
Scenario One hour 313 64 63 75 29
II shift

Scenario One hour 313 64 25 25 37

III shift

Average penetration: The generated wind power compared to the total electric load.
Electric load served by wind: The share of the electric load covered by wind power.
Diesel savings inc. thermal: The total diesel savings for the scenario.
Heating/tap water load served by wind: The share of tap water (sc II) or heating (sc III) load
covered by the wind power.
Useful wind energy: Share of the generated wind power utilized in the system.

4.7 Cost of Energy

The economic analysis in this section will not be a complete investment analysis. Configuring
a complex economical model requires a lot of work, especially collecting the cost parameters.
Local conditions can have great impact on costs, for instance related to the installation of the
windmill, the requirement for grid investments and the transport cost of components. Few
components can be used “off the shelf”, but have to be adjusted and possibly tailor made for
each location.
Instead, an attempt will be made to compare the different scenarios to a pure Wind-Diesel
setup. The extra energy yield that can be obtained from a DHT system will be compared to
the required investment cost for the necessary additional system components, such as the
DHT units and the control and communications system. The goal is to find the COE of the
utilized thermal energy that would otherwise go to waste.

53
4.7.1 Energy savings
When calculating the energy savings, the scenarios are compared to a reference system where
the electric load is covered by diesel generators and tap water/space heating load by fuel oil
combustion (at 90% efficiency).

Scenario I: Wind-Diesel
Reference system: Diesel: 421780 litres
Scenario I: Diesel: 189395 litres
Savings: Diesel: 232385 litres

Scenario II: Wind-Diesel-DHT (tap water)

Reference system: Diesel: 421780 litres Fuel Oil: 42464 litres
Scenario II: Diesel: 161180 litres Fuel Oil: 11441 litres
Savings: Diesel: 260600 litres Fuel Oil: 31023 litres

Scenario III: Wind-Diesel-DHT (space heating)

Reference system: Diesel: 421780 litres Fuel Oil: 317052 litres
Scenario III: Diesel: 161180 litres Fuel Oil: 239384 litres
Savings: Diesel: 260600 litres Fuel Oil: 77668 litres

4.7.2 Energy cost

The energy costs are gathered from Faroese companies in June 2006.
Electricity: The electricity price on the Faroe Islands is general for all consumers, but scales
with the consumption. For consumption between 0-10000 kWh per year, which applies for the
large majority of the Faroese households, the price is 1.11 DKR/kWh, corresponding to 0.15
€/kWh (SEV 2006).
Diesel: The only available diesel price was the price per litre delivered from a service station.
There will most likely be discounts with larger purchases, but since the rate was unknown, the
price was reduced from 6.83 DKR/litre to 5.00 DKR/litre (0.682 €/litre).
Fuel oil: The fuel oil price, derived from the price per 1000 litres and delivered from a tanker,
is 5.813 DKR/litre, corresponding to 0.793 €/litre (Shell 2006).

4.7.3 Investment cost

It will be assumed that the only components needed to expand a wind-diesel system to a
wind-diesel-DHT system are (1) the DHTs, (2) the communication carriers, (3) the central
load commander and (4) the distributed load controller. Estimating the cost of these
components can be complicated, because these systems are not widely commercially
available. The best source was found to be (Johnson et al, 2002), where the cost of the
communication hardware alone was estimated to approximately $1200 (938 €) per load
control unit, the control hardware to 50% of the total cost of the heating system, and the
thermal storage units to $800 - $1500 (626-1173 €) per household. The compared storage
units are not DHTs but ETSs (electric thermal units), but the DHT price is assumed to lie in
the same range.
Scenario III was found to perform better and better with a decreased storage size. It might
even be possible to skip the DHT altogether and install an electric heating element directly in

54
the already existing heating unit. This will further decrease investment cost. However, for the
economic analysis performed in this chapter, both scenarios II and III will include a DHT
unit.
The cost of the control hardware was claimed to be 50% of the total cost of the heating system
for a 16 unit system (Johnson et al, 2002). For the Nólsoy case, with over 100 units, this
percentage will most likely be lower. Both because there are more units to spread the cost on,
but also because the price of integrated circuits, controllers and communication equipment
have been steadily decreasing. For the economic analysis, the cost will be reduced to 25% of
the total heating system cost.
The total investment costs per household will be set to 2640 € with a lifetime of 20 years, and
is assumed to be on the high end of the cost estimate range. The interest rate is set to 6%.

4.7.4 Results
Priority Annual Total annual Annual tap Payback time COE/kWh
System control diesel tap water/heating tap thermal
savings [€] water/heating revenue pr water/heating energy, 20
revenue [€] household [€] system [years] years
lifetime [€]

Scenario - 158487 - - - -
I
Scenario One hour 177729 24601 241 19 0.075
II shift

Scenario One hour 177729 61591 604 6 0.030

III shift

It is important to specify that this is only a rough estimate. The biggest elements of
uncertainty is the investment and installation cost. Secondly the economical model is
simplified, and does not take into account maintenance cost and changes in energy price. Still,
the results are interesting. Both scenario II and III turned out to be profitable compared to a
pure fossil fuelled thermal system. The reason scenario III performed so much better than
scenario II was that it didn’t require any additional investments. In fact, scenario III would
most likely be cheaper, because the simulations show that it might not require any additional
storage elements to perform well.
One element that has not been included in the cost analysis is the increased windmill size of
scenario II and III. The wind energy that was left after the electrical load had been covered
was considered to have no alternative cost, meaning it would otherwise go to waste. In
practice, the installation cost of a windmill is often more than 50% of the total cost, meaning
it is common to oversize the WECS. However, some of this additional cost should be included
in the thermal cost analysis.
The profitability of a pure wind-diesel system was not thoroughly investigated, but appears to
have a good potential. The reduced annual diesel consumption corresponds to over 150 000 €,
enough to maintain an investment of 1 800 000 € over a 20-years period at 6% interest.

55
5 Conclusions and Recommendations
The goal of this thesis was to investigate a renewable energy system where excess wind
power was transformed and stored as thermal energy.
The west-Nordic region has numerous communities that are not connected to the central grid,
and relies on their own electricity production. The electricity and heat generation is to a great
extent based on fossil fuel. With the increasing focus on renewable energy and a steady rising
oil price, there is a growing interest for renewable energy solutions both among the politicians
and the local communities.

Energy consumption
The first part of the project was devoted to local energy planning of the community of Nólsoy.
The yearly consumption and consumption profile of electricity and oil were identified. This
included a comprehensive survey on buildings, utilities and energy consumption, handed out
to every household on the island. The average energy consumption for a household on Nólsoy
was estimated to 8395 kWh electricity and 3125 litres of fuel oil (28125 kWh gross).

Simulation results
Achieving a high share of renewable energy in a stand-alone system can be challenging. Wind
power production is very fluctuating, and an energy storage element can greatly increase the
amount of utilized energy. However, large scale storage of electricity is expensive and
technically challenging. The idea investigated in this thesis was to convert the excess wind
power to thermal energy in decentralized domestic hot water tanks, DHTs. DHTs are
substantially cheaper than electricity storage devices such as batteries and hydrogen systems,
and give the possibility of substituting some of the large amounts fossil fuel combusted for tap
water and heating with renewable energy.
Three different energy scenarios based on a wind-diesel system were constructed and
simulated in a transient simulation program (TRNSYS).
Scenario I was a classic wind-diesel system, and was mainly used as a reference. The good
wind condition in the region, coupled with a good match between the wind- and the load
profile resulted in a halving of the diesel consumption compared to the reference system with
diesel generators. However, more than 50% of the wind energy had to be dumped.
In scenario II, the excess wind power was diverted to a thermal load controller which in turn
controlled a number of distributed DHTs. These units were switched on and off to match the
excess power according to a generated priority routine.
The simulation was run while varying tank parameters to study the effect on system
performance. In general, increasing the storage tank size while holding the rated heating
power within an interval from 750 – 1500 W gave the best results, while changing the tank
temperature did not affect the system substantially. The overall best result from both a
practical and a performance point of view was achieved with a tank size of 1 m2 a heating
element of 1000 W. At this setting the DHTs covered 75% of the tap water load, reducing the
total oil consumption for tap water heating by 31023 litres.
Scenario III was similar to scenario II in layout, but instead of heating tap water the excess
wind power was used to cover the space heating demand. Unlike scenario II, a small storage
size and an increased heating element yielded the best results, more precisely 0.05 m3 and
1500 W. At the most, 25% of the heat demand was covered by excess wind power, saving
77668 litres of oil.

56
The DHTs in scenario II and III were originally controlled by a temperature priority
controller, meaning that the tank with the lowest water temperature was first in line to receive
excess power. A disadvantage with this system is that it requires a two-way communication
between the load controller and the distributed loads, increasing investment cost. The
scenarios were therefore simulated with a simplified control system where the priority levels
for each DHT were changed at a regular basis, something that only requires a one-way
communication system. The results showed that this had this had negible effect on the system
performance.
The cost of energy for the thermal energy supplied from the DHTs was estimated using the
investment cost for the storage tanks, load controllers and communication utilities. Scenario
III turned out to be most economical with a COE of 0.03 €/kWh, compared to 0.075 €/kWh
for scenario II, with a payback time of 6 years and 19 years accordingly. Fuel oil is currently
priced at around 0.079 €/kWh. The cost of energy figures are only approximations, but it
seems clear that scenario III is the most economically viable one.
A pure wind-diesel system like scenario I can reduce the annual diesel consumption with an
amount corresponding to over 150 000 € in fuel cost, enough to maintain an investment of
1 800 000 € over a 20-years period at 6% interest. Utilizing excess wind power in distributed
domestic hot water tanks can help reduce the consumption of fossil fuel further, and at the
same time be competitive on price compared to fuel oil.

Recommendations for further work

Simulation model
• One disadvantage with the simulation model was that only four DHTs were simulated
in detail. This affected the amount of excess wind power that could be utilized, since
the possibility of following the excess power closer is much greater with a higher
resolution. One possibility for including more DHTs could be to write a new TRNSYS
type in FORTRAN that can automatically control a larger amount of DHTs without
demanding a manual configuration.
• The time detail could probably have been lowered to improve the simulation accuracy,
but this requires a higher resolution on the input data for a full benefit.
• Include a safety dump load as a short-time buffer that can absorb fluctuations in the
power.
• Use a complete yearly correlated wind profile in the simulations when it becomes
available.

General
• Initialize a load measurement on the sub sea power cable in order to get better data for
both the yearly and the 24 hours load curve.
• Monitor fuel oil level for a number of households in order to get a better yearly load
profile for space heating (this has been started by Bjarti Thomsen in the spring 2006).
• Calculate the maximum heating power demand for the households using building data
from the questionnaire.

57
6 Bibliography
• Hanssen, S. O. & Thue, J. V. & Skarstein, Ø. & Gjerstad, F. O. & Novakovic, V.
(1996), Enøk I bygninger – Effektiv energibruk, Universitetsforlaget a/s.
• Larsen, B. M. & Nesbakken, R. (2005), Formålsfordeling av husholdningenes
elektrisitetsforbruk i 2001, Statistisk Sentralbyrå.
• Sørensen, S. (1977), Energy consumption for hot water supply, CIB Commission W
62.
• Cappelen, J. & Laursen, E. V. (2004), The climate of the Faroe Islands – with
climatological standard normals, 1961-1990, Danish Meteorological Institute.
• Lemgart, M.-L. & Ulleberg, Ø. (2005), Muligheter for fornybare energisystemer og
hydrogenteknologi i vest-norden, ECON & IFE.
• SEV (2006), The Faroe Islands’ public energy company, personal contact , web page
(Http://www.sev.fo).
• Shell (2006), web page (Http://www.shell.fo)
• Cotrell, J. & Pratt, W.(2003), Modeling the feasibility of using fuel cells and hydrogen
internal combustion engines in remote renewable energy systems, presented to the
American Wind Energy Association’s 2003 WINDPOWER Conference, Austin,
Texas, National Renewable Energy Laboratory
• Drouilhet, S. & Shirazi, M. (2002), Alaska High-Penetration Wind-Diesel Hybrid
Power System: Theory of Operation. NREL/TP-500-31755. Golden, CO: National
Renewable Energy Laboratory, May 2002.
• Johnson, C. & Abdulwahid, U. & Manwell, J.F. & Rogers, A. (2002), Design and
modelling of dispatchable heat storage in wind/diesel systems, Renewable Energy
Research Laboratory, Dept. of Mechanical and Industrial Engineering, University of
Massachusetts, Amherst, MA 01003.
• Hunter, R. & Elliot, G. (1994), Wind-Diesel Systems: A Guide to the Technology and
its Implementation, Cambridge, Cambridge University Press.
• Drouilhet, S.M. (1999), Power Flow Management in a Wind-Diesel Hybrid Power
System with Short-Term Energy Storage, Windpower ’99, Burlington, VT, AWEA.
• Strømmen, K. (2005), Decentralized energy supply based on renewable energy
sources, Department of Electrical Engineering, NTNU.
• Hagstova Føroya (2005), Faroe Islands’ Statistics Department, web page,
(Http://www.hagstova.fo).
• SSB, Statistisk Sentralbyrå (2006), web page, (Http://www.ssb.no).
• UNSD, United Nations Statistics Division (2005), Dioxyde de carbone (CO2),
émissions en tonnes de CO2 par personne (CDIAC), web page,
(Http://millenniumindicators.un.org/unsd/mifre/mi\_series\_results.asp?rowID=751\&
fID=r15\&cgID=).
• Züttel, A., Hydrogen storage methods and materials, University of Fribourg

58
• Skrecc, Sout Kentucky Rural Electric Cooperative Corporation (2006), web page,
(Http://www.skrecc.com/ets.htm).
• CEC, Clark Energy Cooperative (2006), web page,
(Http://www.clarkenergy.com/ets.phtml).
• Wikipedia (2006), web page,
(Http://en.wikipedia.org/wiki/Power_line_communication),
(Http://en.wikipedia.org/wiki/Pumped-storage_hydroelectricity).
• Rekstad, J. (2000), Utnyttelse av solenergi til oppvarming, Faglærer-seminar NTNU
den 4.4 2000.
• Jacobsen, A.(2005), Master Thesis
• HYDROGEMS (2005), Institutt for Energiforskning, web page,
(Http://www.hydrogems.no).
•

59
7 Appendix
7.1 Appendix I - Simulation configuration
7.1.1 System settings
All changes from the basic configurations are marked as bold.

Scenario I - Wind-Diesel: Basic configuration

Value Denom Description
Pmax_EL 265000 KW Maximum (hourly) electric power load
Pmax_Heat 500000 KW Maximum (hourly) thermic power load
Prated_DEGS 100 KW Rated power of the diesel generator
Pmax_DEGS Prated_DEGS*1.2
Pmin_DEGS Prated_DEGS*0.4
Nmin_DEGS 1 Lowest number of diesel generators allowed
Nmax_DEGS 5 Highest number of diesel generators allowed

Scenario II - Wind-Diesel-DHT: Basic configuration

Value Denom Description

T_tap 65 C The desired temperature for tap water
T_source 10 C Temperature of the water source
Etap_year 3794*3600*Nhouseh_DHT KJ Total yearly energy consumption for tap water
PdumpUnitRatio 1/102 The share of the excess wind energy available per DHT
P_DHT 2000 W Rated power of the DHT
P_DHT_hr P_DHT*3.600 KWh
V_DHT 1 m^3 Tank volume of the DHT
Nhouseh_DHT 1 Number of households per DHT
Pmax_EL 265000 KW Maximum (hourly) electric power load
Tset_DHT 85 C Set temperature for DHT
DB_DHT 5 C Temperature deadband for DHT
Prated_DEGS 100 KW Rated power of the diesel generator
Pmax_DEGS Prated_DEGS*1.2
Pmin_DEGS Prated_DEGS*0.4
Nmin_DEGS 1 Lowest number of diesel generators allowed
Nmax_DEGS 5 Highest number of diesel generators allowed
Cp_water 4.186 KJ/kg
(Etap_year/8760)/(Cp_water*
Qavg_hour (T_tap - T_source)) litre/hr Average tap water consumption per hour
T_envi 15 C Environment temperature
T_init 65 C Initial temperature of water in DHT

Scenario II - Wind-Diesel-DHT: Reference system

Value Denom Description

P_DHT 0 W Rated power of the DHT
T_envi 10 C Environment temperature

60
T_init 10 C Initial temperature of water in DHT

Scenario II - Wind-Diesel-DHT: Default settings

Value Denom Description
T_envi 10 C Environment temperature
T_init 10 C Initial temperature of water in DHT
Hnode_DHT 1.0839/Nnode_DHT M Height of the nodes in the DHT

Scenario II - Wind-Diesel-DHT: Storage tank volume

Value Denom Description
V_DHT 0.25-3 m^3 Tank volume of the DHT
T_envi 10 C Environment temperature
T_init 10 C Initial temperature of water in DHT
Nnodes_DHT 15 Number of temperature nodes in the DHT
Hnode_DHT (0.68-1.56)/Nnode_DHT M Height of the nodes in the DHT

Scenario II - Wind-Diesel-DHT: Maximum heating rate of DHT

Value Denom Description
P_DHT 500-5000 W Rated power of the DHT
T_envi 10 C Environment temperature
T_init 10 C Initial temperature of water in DHT
Nnodes_DHT 15 Number of temperature nodes in the DHT
Hnode_DHT 1.08/Nnode_DHT M Height of the nodes in the DHT

Scenario II - Wind-Diesel-DHT: Temperature settings

Value Denom Description
Tset_DHT 65-95 C Set temperature for DHT
DB_DHT 2.5-5-7.5 C Temperature deadband for DHT
Prated_DEGS 1000 KW Rated power of the diesel generator
T_envi 10 C Environment temperature
T_init 10 C Initial temperature of water in DHT
Nnodes_DHT 15 Number of temperature nodes in the DHT
Hnode_DHT 1.08/Nnode_DHT M Height of the nodes in the DHT

Scenario III - Wind-Diesel-DHT (heating): Reference system

Value Denom Description

PdumpUnitRatio 1/102 The share of the excess wind energy available per DHT
P_DHT 0 W Rated power of the DHT
P_DHT_hr P_DHT*3.600 KWh
V_DHT 1 m^3 Tank volume of the DHT
Nhouseh_DHT 1 Number of households per DHT
Pmax_EL 265000 KW Maximum (hourly) electric power load
Pmax_Heat 500000
Tset_DHT 85 C Set temperature for DHT

61
DB_DHT 5 C Temperature deadband for DHT
Prated_DEGS 100 KW Rated power of the diesel generator
Pmax_DEGS Prated_DEGS*1.2
Pmin_DEGS Prated_DEGS*0.4
Nmin_DEGS 1 Lowest number of diesel generators allowed
Nmax_DEGS 5 Highest number of diesel generators allowed
Cp_water 4.186 KJ/kg
(Etap_year/8760)/(Cp_water* Average tap water consumption per hour
Qavg_hour (T_tap - T_source)) litre/hr
T_envi 10 C Environment temperature
T_init 20 C Initial temperature of water in DHT
Nnodes_DHT 15 Number of temperature nodes in the DHT
Hnode_DHT 1.08/Nnode_DHT M Height of the nodes in the DHT
T_radiator 65 C Temperature of the water going to the radiators
Eheat_year 28125*3600*Nhouseh_DHT KJ The total heating demand per household per year
Pavg_hour Eheat_year/8760 KJ Average heating demand per hour
Qheatavg_hour 200 litre/hour Flow rate of water in the radiator circuit
T_indoor 20 C Indoor temperature

Scenario III - Wind-Diesel-DHT (heating): Default settings

Value Denom Description

P_DHT 1000 W Rated power of the DHT

Scenario III - Wind-Diesel-DHT (heating): Storage tank volume

Value Denom Description
P_DHT 1000 W Rated power of the DHT
V_DHT 0.05-3 m^3 Tank volume of the DHT

Scenario III - Wind-Diesel-DHT (heating): Maximum heating rate of DHT

Value Denom Description
P_DHT 500-6000 W Rated power of the DHT
V_DHT 0.5 m^3 Tank volume of the DHT

62
7.2 Appendix II – Questionnaire
Note that this is not the original design of the questionnaire, only the template for adding data
to the database. The questions are however identical (translated to English).

63
64