1. No category

Depar tment of Applied Physics
Modeling and optimization
of urban energy systems for
large-scale integration of
variable renewable energy
generation
Jani Mikkola
D O C TO R A L
DISSERTATIONS
Aalto University publication series
DOCTORAL DISSERTATIONS 16/2017
Modeling and optimization of urban
energy systems for large-scale
integration of variable renewable
energy generation
Jani Mikkola
A doctoral dissertation completed for the degree of Doctor of
Science (Technology) to be defended, with the permission of the
Aalto University School of Science, at a public examination held at
the lecture hall K216 of the school on 24th of February 2017 at 12
noon.
Aalto University
School of Science
Department of Applied Physics
New Energy Technologies
Supervising professor
Professor Peter Lund, Aalto University, Finland
Thesis advisor
Professor Peter Lund, Aalto University, Finland
Preliminary examiners
Professor Lennart Söder, KTH Royal Institute of Technology, Sweden
Professor Torjus Folsland Bolkesjø, NMBU Norwegian University of Life Sciences, Norway
Opponent
Professor Reinhard Haas, TU Wien, Austria
Aalto University publication series
DOCTORAL DISSERTATIONS 16/2017
© Jani Mikkola
ISBN 978-952-60-7266-1 (printed)
ISBN 978-952-60-7265-4 (pdf)
ISSN-L 1799-4934
ISSN 1799-4934 (printed)
ISSN 1799-4942 (pdf)
http://urn.fi/URN:ISBN:978-952-60-7265-4
Unigrafia Oy
Helsinki 2017
Finland
Ab s t ra c t
Aalto University, P.O. Box 11000, FI-00076 Aalto www.aalto.fi
Author
Jani Mikkola
Name of the doctoral dissertation
Modeling and optimization of urban energy systems for large-scale integration of variable
renewable energy generation
P u b l i s h e r School of Science
U n i t Department of Applied Physics
S e r i e s Aalto University publication series DOCTORAL DISSERTATIONS 16/2017
F i e l d o f r e s e a r c h Energy sciences
M a n u s c r i p t s u b m i t t e d 12 September 2016
D a t e o f t h e d e f e n c e 24 February 2017
P e r m i s s i o n t o p u b l i s h g r a n t e d ( d a t e ) 1 November 2016
Monograph
Article dissertation
L a n g u a g e English
Essay dissertation
Abstract
The rapid growth of variable renewable energy (VRE), such as wind and solar energy, will increase
the unpredictability and fluctuations in energy systems in general, resulting in a need for increased
fl exibility measures. Simultaneously, global urbanization will strengthen the importance of urban
energy systems in the future. Therefore, it is highly important to fi nd solutions to enable largescale variable renewable energy generation in cities.
This dissertation aims to fi nd ways to increase fl exibility and to enable higher shares of VRE in
urban energy systems. For this purpose, three new computational models were developed in
relation to both electricity and thermal energy (heating and cooling) in cities. The fi rst model
simulates the spatiotemporal energy consumption in cities, the second focuses on distributed
production and cooperation between different energy networks, and the third describes the energy
system operation with existing power plants and large-scale VRE schemes.
Results from Helsinki, Finland, show that power-to-heat (P2H) conversion with electric
resistance or heat pumps can help to increase the wind power production in the city by 4–5 times
(up to 60% of electricity demand) compared to the self-use limit of wind power, in which the focus
is on the electric system only. Part of the increased wind power production would then be used to
cover up to one third of the heat demand. Simultaneously, P2H makes coal very sensitive to price
changes, which could lead to the removal of coal from the city's energy portfolio.
In another case, in Concepción, Chile, the solar capacity of the city is, in theory, large enough
to satisfy all energy demand, but the technical effi ciency and the power grid infrastructure limits
the PV share in electricity to around 24%. With P2H, this could be increased to at least 34% without
overloading the power grid; simultaneously a quarter of the heat demand would be fi lled. These
results demonstrate the importance of system-wide consideration when planning integration of
VRE.
The thesis shows that, by viewing energy systems as a whole instead of focusing on only one
energy form, the fl exibility of the system can be remarkably increased, even without massive
investments in traditional network infrastructure. Through smart design and multi-energy thinking
approaches, variable renewable energy sources can be turned into a main-stream option for energy
production in urban areas.
K e y w o r d s urban energy systems, energy system modeling, renewable energy, optimization,
energy fl exibility
I S B N ( p r i n t e d ) 978-952-60-7266-1
I S B N ( p d f ) 978-952-60-7265-4
I S S N - L 1799-4934
I S S N ( p r i n t e d ) 1799-4934
I S S N ( p d f ) 1799-4942
L o c a t i o n o f p u b l i s h e r Helsinki
L o c a t i o n o f p r i n t i n g Helsinki Y e a r 2017
Pages 170
u r n http://urn.fi /URN:ISBN: 978-952-60-7265-4
Tiivistelmä
Aalto-yliopisto, PL 11000, 00076 Aalto www.aalto.fi
Tekijä
Jani Mikkola
Väitöskirjan nimi
Kaupunkien energiajärjestelmien mallinnus ja optimointi laajamittaisen vaihtelevan uusiutuvan
energian lisäämiseksi
J u l k a i s i j a Perustieteiden korkeakoulu
Y k s i k k ö Teknillisen fysiikan laitos
S a r j a Aalto University publication series DOCTORAL DISSERTATIONS 16/2017
T u t k i m u s a l a Energiatieteet
K ä s i k i r j o i t u k s e n p v m 12.09.2016
V ä i t ö s p ä i v ä 24.02.2017
J u l k a i s u l u v a n m y ö n t ä m i s p ä i v ä 01.11.2016
Monografia
Artikkeliväitöskirja
K i e l i Englanti
Esseeväitöskirja
Tiivistelmä
Vaihtelevan uusiutuvan energian (VRE), esimerkiksi tuuli- ja aurinkoenergian, nopea yleistyminen
kasvattaa maailmanlaajuisesti epävarmuutta ja tuotannon heilahduksia energiajärjestelmissä.
Tämän vuoksi tarvitaan uusia keinoja järjestelmien joustavuuden lisäämiseksi. Samalla
kaupunkien kasvu ja globaali urbanisaatio nostavat kaupunkiympäristöjen merkitystä
energiakysymyksissä. Onkin erittäin tärkeää keksiä ratkaisuja, joilla laajan mittakaavan VREtuotanto kaupungeissa mahdollistetaan.
Tässä väitöskirjassa etistään kaupunkeihin sopivia keinoja, joilla energiajärjestelmien
joustavuutta voidaan lisätä ja uusiutuvan energian osuutta kasvattaa. Tarkoitusta varten luotiin
kolme uutta tietokonemallia, jotka ottavat huomioon sekä sähkön että termisen energian
(lämmityksen ja jäähdytyksen). Ensimmäinen malli simuloi kaupunkien ajallista ja paikallista
energiankulutusta, toinen keskittyy hajautettuun tuotantoon ja eri energiaverkkojen
yhteiskäyttöön, ja kolmannella mallinnetaan energiajärjestelmän toimintaa, kun mukana on
vanhoja voimalaitoksia ja paljon vaihtelevaa uusiutuvaa tuotantoa.
H e l s i ng is t ä s a ad u t t u l o ks e t o s o i t t ava t , e t t ä s äh kö – l ä mp ö - ko nve rs i o ( P2H ) j o ko s äh kö vas t u ks i l l a
tai lämpöpumpuilla auttaa nostamaan tuulivoiman tuotannon määrää kaupungissa jopa
nelin–viisinkertaiseksi (jopa 60 % sähköstä) verrattuna perinteiseen lähestymistapaan, jossa
keskitytään yksinomaan sähköjärjestelmän toimintaan. Osa lisääntyneestä tuulituotannosta
käytetään kattamaan jopa kolmannes kaupungin lämmöntarpeesta. Samalla P2H tekee hiilivoiman
erittäin herkäksi hinnanvaihteluiden suhteen, mikä saattaa johtaa luonnostaan kokonaan hiilestä
luopumiseen.
Chilen Concepciónissa kaupungin aurinkoenergian potentiaali riittäisi kattamaan kaupungin
koko energiantarpeen, mutta tekninen suorituskyky ja verkkoinfrastruktuuri rajoittavat
aurinkosähkön osuuden noin 24 %:iin. P2H:n avulla osuus voidaan nostaa vähintään 34 %:iin
ylikuormittamatta silti verkkoa. Samalla tuotettaisiin neljännes kaupungin lämmöntarpeesta.
Tulokset osoittavat järjestelmätason tarkastelun tarpeellisuuden uusiutuvan energian
suunnittelussa.
Väitöskirja todistaa, että tarkastelemalla energiajärjestelmää kokonaisuutena (yksittäisen
energiamuodon sijasta) sen joustavuutta voidaan lisätä ilman valtavia perinteisiä
verkkoinvestointeja. Älykkäällä suunnittelulla ja monienergia-ajattelulla vaihtelevat uusiutuvat
energiamuodot voidaan nostaa kaupunkiympäristöjen tärkeimmäksi energiamuodoksi.
A v a i n s a n a t kaup unkie n e ne rgiajärje ste lmät, e ne rg iajärje ste lmie n mallinnus, uusiutuva e ne rg ia,
optimointi, energiajoustavuus
I S B N ( p a i n e t t u ) 978-952-60-7266-1
I S B N ( p d f ) 978-952-60-7265-4
I S S N - L 1799-4934
I S S N ( p a i n e t t u ) 1799-4934
I S S N ( p d f ) 1799-4942
J u l k a i s u p a i k k a Helsinki
P a i n o p a i k k a Helsinki
Sivumäärä 170
u r n http://urn.fi /URN:ISBN: 978-952-60-7265-4
V u o s i 2017
Preface
This thesis is based on research carried out in the New Energy Technologies
Group of the Department of Applied Physics, Aalto University School of Science,
under the supervision of Professor Peter Lund. Funding for the work was provided by Fortum Foundation, Academy of Finland, and the Finnish Funding
Agency for Innovation (Tekes). The financial support is greatly appreciated.
I want to thank my instructor and supervisor, Professor Peter Lund for providing valuable support, guidance and inspiration through my whole academic career, and also for making it possible for me to work with a topic I found interesting and important. In the beginning of my research career, I was instructed
by Dr. Mikko Mikkola and Rami Niemi, I want to express my gratitude also to
them.
I wish to thank the colleagues who helped me and took part in the research,
Juuso Lindgren, Jyri Salpakari, and Joel Ypyä. I want to acknowledge also my
Chilean colleagues Paulina Wegertseder and Rodrigo García. Furthermore, I
want to thank our whole research group, everyone else I have worked with, and
all my friends.
Finally, I want to express my gratitude to my wife and my whole family for all
the love and support throughout my life.
Otaniemi, Espoo, January 17, 2017
Jani Mikkola
1
Contents
Preface ...................................................................................................... 1
List of abbreviations and symbols ........................................................... 5
List of publications .................................................................................. 8
Author’s contribution .............................................................................. 9
1.
Introduction................................................................................. 11
1.1
Background .............................................................................. 11
1.2
Objectives and scope ................................................................ 12
1.3
Thesis outline ........................................................................... 13
2.
Urban energy systems ................................................................. 15
2.1
Electric power systems............................................................. 15
2.2
Heating and cooling ................................................................. 16
3.
Energy system flexibility (Publication I) ..................................... 19
3.1
Flexibility challenge of renewable integration ......................... 19
3.2
Flexibility technologies ............................................................ 19
3.2.1
Supply-side flexibility and grid services............................... 19
3.2.2
Energy storage ..................................................................... 20
3.2.3
Power-to-X ........................................................................... 21
3.2.4
Demand-side management ................................................. 22
3.2.5
Novel grid infrastructure ..................................................... 22
4.
Spatiotemporal consumption profiles (Publication II) .............. 25
4.1
Input data................................................................................ 25
4.2
Formation of spatiotemporal load profiles ............................. 26
4.2.1
Scaling base profiles to match the city demand .................. 26
4.2.2
Sub-class division ................................................................ 26
4.2.3
Spatial division .................................................................... 28
5.
Multi-carrier energy networks (Publication III) ........................ 29
5.1
Structure of the multi-carrier energy systems model ............. 29
5.2
Energy balance in nodes and flows in network ....................... 31
5.3
Energy system operation in simulations ................................ 32
3
6.
Optimal use of an energy system under large-scale VRE
(Publication V) ............................................................................ 35
6.1
Model principles ...................................................................... 35
6.2
Input and output .....................................................................36
6.2.1
VRE production, consumption, and technical parameters .36
6.2.2
Economic parameters ......................................................... 38
6.2.3
Output ................................................................................. 38
6.3
Optimization of energy system operation ...............................39
6.3.1
Objective function ................................................................39
6.3.2
Operation principles of an energy system – boundary
conditions.............................................................................39
7.
Results for Helsinki ..................................................................... 41
7.1
Current energy system in Helsinki .......................................... 41
7.2
Spatiotemporal consumption profiles (Publication II) ...........43
7.2.1
Input data.............................................................................43
7.2.2
Spatiotemporal consumption profiles – results ..................44
7.3
P2H conversion enables more VRE (Publications III and IV) 46
7.3.1
Wind power with P2H and thermal storage ........................ 47
7.3.2
Effect of network structure ................................................. 48
7.4
Energy system operation with CHP plants and large-scale
wind power (Publication V) .................................................... 50
7.4.1
Parameters for Helsinki energy system .............................. 50
7.4.2
Effects of wind power and P2H on plant operation............. 51
8.
Results for other cities................................................................. 55
8.1
Shanghai, China....................................................................... 55
8.1.1
Spatiotemporal profiles with limited input data
(Publication II)..................................................................... 55
8.1.2
Large-scale photovoltaics together with microtrigeneration (Publication III) ............................................. 57
8.2
9.
Concepción, Chile (Publication VI) .........................................58
8.2.1
Electricity consumption, city structure, and the power
grid .......................................................................................58
8.2.2
Results.................................................................................. 59
Conclusions ................................................................................. 61
References .............................................................................................. 65
Publications ............................................................................................ 71
4
List of abbreviations and symbols
Abbreviations
CAES
compressed-air energy storage
CHP
combined heat and power
CO2
carbon dioxide
COP
coefficient of performance
DH
district heating
DSM
demand-side management
GHG
greenhouse gas
HVDC
high-voltage direct current
MILP
mixed-integer linear programming
O&M
operations and maintenance
P2G
power-to-gas
P2H
power-to-heat
P2T
power-to-thermal
P2X
power-to-X
PHES
pumped hydropower energy storage
PV
photovoltaics
TSO
transmission system operator
V2G
vehicle-to-grid
VRE
variable renewable energy
Symbols
‫ܣ‬
floor area density (m2/km2)
‫ܤ‬
energy balance (MW)
5
ܾ
relative base load profile (–)
ܿ
energy converted to other forms (MW)
݀
density of urban structure (unit/km2)
݀଴
maximum density (unit/km2)
ds
downstream nodes (–)
݂
energy power flow (MW)
‫ܫ‬
energy intensity (MWh/m2)
‫ܭ‬
constant minimum density of urban structure (unit/km2)
ܲ
energy production (MW or MWh)
ܳ
energy consumption (MW or MWh)
‫ݍ‬
energy allocated for conversion to other energy forms (MW)
‫ݏ‬
energy charged to or discharged from storage (MW)
‫ݐ‬
time, time step (min, h)
‫ݔ‬
spatial coordinate (km), or distance from the city center (km)
‫ݔ‬଴
distance of the maximum density from the city center (km)
‫ݕ‬
spatial coordinate (km)
‫ݖ‬
energy curtailment (MW)
ߙ
width parameter of density distribution (1/km2)
ߟ
efficiency (–)
ߣ
scaling factor (–)
Subscripts and superscripts
ܿ
cooling
city
city total
݁
electricity
݂
fuel
݄
heat
݅
energy consumer class
in
incoming
݆
energy consumer subclass
݇
energy form
6
Introduction
݈
line of an energy network
݊
node
out
outgoing
‫ݐ‬
time step
tot
total
ߢ
other energy forms (than the one in question)
7
List of publications
This doctoral dissertation consists of a summary and of the following publications which are referred to in the text by their Roman numerals.
I. Lund, Peter; Lindgren, Juuso; Mikkola, Jani; Salpakari, Jyri. Review of energy system flexibility measures to enable high levels of variable renewable
electricity. Renewable & Sustainable Energy Reviews 45, 785–807 (2015).
II. Mikkola, Jani; Lund, Peter. Models for generating place and time dependent urban energy demand profiles. Applied Energy 130, 256–264 (2014).
DOI: 10.1016/j.apenergy.2014.05.039
III. Niemi, Rami; Mikkola, Jani; Lund, Peter. 2012. Urban energy systems
with smart multi-carrier energy networks and renewable energy generation.
Renewable Energy 48, 524–536 (2012). DOI: 10.1016/j.renene.2012.05.017
IV. Lund, Peter; Mikkola, Jani; Ypyä, Joel. Smart energy system design for
large clean power schemes in urban areas. Journal of Cleaner Production
103, 437–445 (2015). DOI: 10.1016/j.jclepro.2014.06.005
V. Mikkola, Jani; Lund, Peter. Modeling flexibility and optimal use of existing
power plants with large-scale variable renewable power schemes. Energy
112, 364–375 (2016). DOI: 10.1016/j.energy.06.082
VI. Wegertseder, Paulina; Lund, Peter; Mikkola, Jani; García Alvarado, Rodrigo. Combining solar resource mapping and energy system integration
methods for realistic valuation of urban solar energy potential. Solar Energy
135, 325–336 (2016). DOI: 10.1016/j.solener.2016.05.061
8
Author’s contribution
Publication I: Review of energy system flexibility measures to enable high
levels of variable renewable electricity
The author was responsible for the Chapters 6 (Supply-side flexibility), 7 (Advanced technologies, except Subchapter 7.3) and 8 (Infrastructure). He conducted the literature review of these topics and was the main author of these
chapters.
Publication II: Models for generating place and time dependent urban energy demand profiles
The author developed the model presented in the paper. He did the modeling
work and calculations. The author was the main author of this publication.
Publication III: Urban energy systems with smart multi-carrier energy
networks and renewable energy generation
The author developed the idea of the paper with the co-authors. He conducted
the modeling work and calculations. He wrote part of the text.
Publication IV: Smart energy system design for large clean power schemes
in urban areas
The author used the model he had developed to calculate spatial load profiles
and conducted spatiotemporal analyses. He took part in writing the paper.
Publication V: Modeling flexibility and optimal use of existing power plants
with large-scale variable renewable power schemes
The author developed the optimization model. He wrote the computational
codes, and calculated and analyzed the results. He was the main author of the
paper.
Publication VI: Combining solar resource mapping and energy system integration methods for realistic valuation of urban solar energy potential
The author developed a sub-model and run the simulations about energy system integration. He wrote the corresponding part of the text.
9
1. Introduction
1.1
Background
The reduction of greenhouse gas (GHG) emissions is the main option for mitigating global climate change. Approximately 60–70% of the global GHG emissions stem either directly or indirectly from energy production [1,2]. Therefore,
massive amounts of clean and renewable energy production are now being installed in the world. Several studies predict that a remarkable share of the new
energy production will be based on variable renewable energy (VRE) sources,
such as wind and solar power [3,4]. The potential of VRE, especially solar energy, is enormous. Each year, the sun radiates over 5,000 times more energy on
the ground than mankind consumes [5]. The use of these technologies will gain
increasing popularity as the technology develops and the price is reduced.
The fluctuating and decentralized production of VRE sources creates new
types of challenges to current energy systems, which are traditionally based in
centralized and controllable combustion power plants and fossil fuels [6]. Local
small VRE installations have not yet caused any major problems in the power
systems, but large-scale VRE scenarios will require a different type of flexibility
from the energy system to balance temporal and spatial differences between energy demand and supply. In electric systems, this balancing is especially critical,
as the systems typically have little inherent inertia and low storage capacity.
The transition from traditional to renewable energy production is a long process and will require the old and new energy systems to cooperate side by side
for decades. Therefore, the traditional plants that may be designed for steady
base-load operation will encounter totally different working conditions when
VRE is introduced on a large scale. This scenario again emphasizes the need for
power system flexibility, as increased fluctuations may decrease the lifetime of
the plants [7] and increase the costs of power system operation [8]. In addition
to increasing system flexibility, it is important to study the optimal operation of
the resulting energy mix of new and old energy systems.
A total of three fourths of all energy is currently consumed in urban areas, and
the share of cities in total GHG emissions is approximately 80% [9]. Furthermore, the role of cities will only increase as urbanization continues in the future
[10]. These facts highlight the importance of solving the challenge of energy system transition, especially in cities. Simultaneously, the progress has been rapid
in distributed energy production technologies suitable for urban environments,
especially solar energy, but also heat pumps, distributed bioenergy, small-scale
wind power, and renewable district heating. Bringing energy production closer
11
Introduction
to energy use and to consumers is an attractive option for the future, as it has
several benefits, including avoiding large and costly energy infrastructures and
involving end-users in energy investments [11].
In urban areas, thermal energy (i.e., heating and cooling) dominates the end
use of energy over electricity. In the residential sector, the share of thermal energy in the end use varies between 50–80% depending on the location, whereas
the share of electricity remains between 15–35% [12]. The difference is a little
smaller in the service sector [13]. Thus, it is highly important to consider the
energy system as a whole and not to limit the scope strictly to electricity alone.
This is particularly important because thermal energy is often produced locally,
as long-distance transportation of heat is not yet technically or economically viable [14].
1.2
Objectives and scope
The purpose of this thesis is to investigate ways of increasing the share of VRE
in the energy system, especially in urban environments. Previous studies on energy system integration of VRE tend to focus on power only. In contrast, this
thesis considers the energy system as a whole, taking into account thermal and
electric energy side by side. Flexibility measures such as power-to-heat conversion (P2H) with thermal storage are given emphasis in this context. The purpose
of looking at the system as a whole is to find new ways to increase the system
flexibility and thus allow larger VRE installations to be done.
Three main topics are covered in this work. First, it is important to recognize
the characteristics of the current energy use; that is, to have data that is as detailed as possible about spatial and temporal energy consumption. Traditionally
urban energy consumption is modeled either spatially or temporally. In this
work, the aim was to combine these dimensions into a single model with a sufficiently detailed resolution. This type of data is needed in matching energy supply and demand but also in optimizing the energy system operation and identifying possibilities for systemic innovations and improvements.
Second, the question of multiple energy carriers and their cooperation and
mutual conversion is addressed. New ways to provide flexibility are reached for
by viewing the urban energy on a system level with different energy forms. For
this purpose a new comprehensive energy-carrier model was developed that includes control functions and network intelligence. Another improvement compared to the existing literature is the possibility of incorporating both macro and
micro levels to the same model. The model makes use of the detailed data from
the spatio-temporal consumption model (e.g., Helsinki, Finland and Shanghai,
China). As an example of the use of this model, the limitations of urban energy
infrastructures on the solar energy potential on a city-level (Concepción, Chile)
was investigated. This was done in a novel way by combining highly-detailed
solar potential mapping with the energy-carrier model developed in this thesis.
Third and finally, the question of the combination of traditional energy systems with large-scale VRE integration is discussed. The thesis considers the
12
Introduction
consequences of large-scale VRE for traditional power plants, as well as the optimal way to operate energy systems under large VRE schemes. There are several dispatch models available. However, many of these are non-linear, which
often leads to long computation times. In addition, the existing models are often
limited to one energy form or few power plants only. The aim here was to cover
the whole urban energy system including both the electricity and thermal energy
sectors, but still keeping the simulation model user friendly and quick.
1.3
Thesis outline
The thesis is structured as follows. The first chapter presents a general background of the topic and presents the objectives of the study. Chapter 2 gives a
concise presentation of energy systems in urban areas. Chapter 3 gives a short
overview of different flexibility methods and technologies in energy systems
(Publication I). In the three following chapters, three different models are presented: Chapter 4 handles the spatiotemporal load profile formation (Publication II), Chapter 5 presents the multi-energy carrier model together with spatial
energy flows (Publication III), and Chapter 6 describes the optimization model
created for energy system operation with large-scale renewable energy schemes
(Publication V). The next two chapters (7 and 8) present the results from the
simulations and optimizations; specifically Chapter 7 focuses on Helsinki, Finland, and Chapter 8 focuses on Shanghai, China and Concepción, Chile. The results are gathered from Publications II–VI. The model developed in Publication
III is used, with minor appropriate modifications, in calculating the results for
Publications IV and VI. Finally, Chapter 9 summarizes and concludes the thesis.
13
2. Urban energy systems
The modern lifestyle is dependent on a constant and secure energy supply. Energy is needed in several forms, such as electricity, heating and cooling, and
fuels. Commonly, electricity is the main focus of the discussion of energy, as it
is needed in most of our modern appliances and in areas from communication
to industrial activities, services and lighting. However, to keep societies working, other forms of energy are also important, and all forms must be considered
in order to understand the whole energy scheme and the challenges posed by
climate change and limited natural resources. Energy systems are ways of producing energy and delivering it to consumers. This chapter is a concise presentation of electric and thermal energy systems, focusing on urban areas.
2.1
Electric power systems
Electric power systems are complex combinations of networks, production
units, transformer stations, end-use connections, and market operators. Their
job is to guarantee a secure and constant power supply to consumers.
Traditionally, electricity is produced in power plants by burning fossil fuels.
In 2013, more than 40% of all electricity in the world was produced by coal,
while 22% was produced by natural gas. Hydropower is the largest renewable
source, with a 16% share [15]. All these energy sources are controllable; thus,
the traditional way of balancing supply with demand has been controlling the
supply side. However, new variable renewable power sources, such as solar and
wind power, are increasing rapidly; as such, it is obvious that new balancing
methods are required [16].
Electrical grids deliver power from the generating plants to consumers. The
plants are connected to energy systems through transformers that change the
voltage of the generated power to the necessary level. The long-distance transmission is typically done by high-voltage lines with a voltage that is typically
between 110–800 kV [17,18]. In cities, the local grid is typically connected to the
national or other larger-area transmission network.
A city network also typically consists of different voltage levels. For example,
the main trunk lines could be high-voltage lines, like the national grid, whereas
the actual distribution grid is typically a low-voltage grid, with voltages below
50 kV, and it is connected to high-voltage lines with transformer stations. Along
15
Urban energy systems
the distribution grid, power stations that are close to actual end users convert
the voltage to the 230 V or 110 V typically used in domestic power plugs [19].
The balance between supply and demand must be maintained in order to keep
voltage and frequency stable in the grid. The voltage drops in a network line
because of current flowing through loads and impedances along the line. To
keep the power system running, the power stations have to regulate the voltage
to stay within a tolerated margin [19]. Distributed generation (e.g., small-scale
solar and wind power) compensates for the voltage drop in the lines, which can
be utilized in power system design [20].
In the future, as distributed and variable generation increases, new kinds of
technological solutions to power system operation will also probably arise. For
example, the role of the urban power grid may change from a one-way channel
of power delivery into a way of balancing spatial differences in supply and demand. This leads to different smart grid concepts, which enable all parties of a
power system to participate actively in the power system operation and balancing [21,22]. Smart grids are discussed further in Chapter 3.2.5.
2.2
Heating and cooling
The need for comfortable room temperatures forms one of the largest shares of
total energy consumption. In urban areas, heating and cooling systems typically
consume more energy than electrical appliances [12]. Heating energy is needed
in two forms: space heating and domestic hot water.
The most common form of heating in the world is local heat production. There
are several options for this form of production, including furnaces or boilers
burning wood, natural gas, propane, or fuel oil. Electricity can also be used to
provide heat through electric resistance or heat pumps [23]. In sunny regions,
solar thermal collectors are common, especially in water heating [24]. Lately,
the role of architecture in heating demand for buildings has strengthened, as
zero-energy buildings have become increasingly popular [25].
An alternative to local production is district heating (DH), which is now available globally. For example, in Europe, the share of district heating is about 10%
of the heating market [26], and it is particularly widely used in Nordic countries
[27]. DH systems are especially suitable for urban areas, whereas the losses of
heat transfer would grow too large in sparsely-populated locations [28].
The idea of DH is to produce heat centrally, usually in large units or CHP
plants, and then to deliver this heat to consumers through hot water or steam
pipelines. These lines are built with two pipes; the first pipe carries the hot water
(or steam) to the customers, where the heat is delivered directly or through heat
exchangers to heating appliances or heat storage. Afterward, the cooled fluid
returns to the heating station inside the other pipeline. The system is operated
by pumps that pressurize the system and keep the fluid moving [28].
The demand for heating varies according to outside temperatures, but the
heating systems must be able to deliver enough heat during the coldest months
of the year. Thus, different types of heat production units are needed in the DH
systems. The base load could be satisfied, for example, by CHP or nuclear power
16
Urban energy systems
plants operating throughout the year, but a reserve capacity is needed for colder
periods. Heat-only boiler stations are commonly used, but heat pumps [29] and
central solar heating [30,31] may also be used to provide heat for DH network
[32].
Easy storage of heat is one advantage of district heating systems. Large hot
water tanks are relatively cheap to build and connect to the heating network,
which also has its own storage capacity. The water temperature inside the pipes
can be increased, thus giving short-term storage capacity for sudden variations
in demand [28].
Similarly, cooling is also needed to keep room temperatures at a comfortable
level. The most common cooling techniques are local small-scale techniques and
appliances, such as ventilation [33], air conditioning [34] or absorption cooling
[35,36]. Nowadays, district cooling systems are also becoming more popular.
Their operation principle is quite similar to district heating, but in reverse: thermal energy is transmitted from the customer to the district cooling system
[37,38].
17
3. Energy system flexibility (Publication
I)
3.1
Flexibility challenge of renewable integration
A power system has to guarantee a continuous energy supply that matches the
fluctuating demand. Traditionally, different types of controllable power plants
have ensured this energy balance. However, with increasing amounts of variable
renewable energy generation, new kinds of flexibility measures are required
[39]. In addition to temporal balance, spatial flexibility is also needed, as the
capacity of distribution and transmission networks may limit the balancing opportunities [40].
Especially in electric power systems, the demand and supply must be in balance at each time point, and the system is able, up to a certain point, to cope
with variations and uncertainty. In electric systems, the flexibility relates close
to grid frequency and voltage control, as well as the security of supply and power
ramping rates [41]. For example, a portfolio of power plants with different response times has traditionally been used to provide the necessary flexibility
[42].
As energy production is becoming more distributed and variable, the old system will need to continue to operate side-by-side with the new technologies for
decades to come. In order to keep the demand and supply in balance in the future, a wide selection of different technologies will be of importance [40].
Publication I presents a broad review of current and future technologies that
could fulfill the flexibility required. The article covers both supply- and demandside options as well as network infrastructure, storage, and energy market questions. Advanced technologies, such as power-to-X or vehicle-to-grid, are also
included.
3.2
3.2.1
Flexibility technologies
Supply-side flexibility and grid services
Traditionally in power systems, the balance between supply and demand has
been maintained by power plants and grid operators. Supply-side flexibility co-
19
Energy system flexibility (Publication I)
vers the measures and technologies to modify the output from the power generation units in order to make the demand and supply match. Grid operators, such
as transmission system operators (TSO), are responsible for keeping the power
grid in balance. In addition to controlling the output of power plants, other
measures, namely grid ancillary services, can also be used for this purpose.
These ancillary services are generic measures to maintain the balance at different time scales, from power quality and regulation with, for example, flywheels and batteries (1 ms – 5 min), to spinning and power plant reserves (5
min – 1 h), to longer (1 h – months) load shifting with large storage facilities
[43,44].
The ability of power plants to respond to demand fluctuations depends on the
type of plants they are. The characteristics of different plant types are listed in
Table 1. Typical examples of base-load power plants include nuclear and coal
power plants. These are designed to operate mainly at constant (preferably
nominal) power, and ramping and shut-down are avoided for economical or
technical reasons. Load-following plants, such as hydropower or gas turbines,
are run according to the balance situation in the grid. They are quick to start up
and ramp the output (with response times ranging from seconds to minutes).
Peaking power plants are used irregularly during high-demand hours [39].
Table 1. Key output and ramping parameters of different power plant technologies [39,45].
Indicator
Typical unit
size (MW)
Efficiency (%)
Steam
Gas
Combi-cy-
turbine
turbine
cle turbine
Engine
Nuclear
600–900
10–300
60–400
1–20
800–1600
40–47
32–38
50–60
45–48
33–37
100–600
100–600
30–180
1–15
(i.e., up to
Start-up and
synchronize
120–3000
time (min)
Ramp-up rate
(%/min)
Minimum output (%)
2 days)
3–6
5–20
3–6
20–25
40
50
40–50
30
30
One option to regulate VRE production is to curtail the output when necessary. Such situations can occur with limited transmission capacity, overly large
and inflexible base-load generation, and oversupply of VRE [46]. Curtailment
can quickly reduce the generation, thus dampening rapid changes in power output, but it can also be used to provide reserve power capacity through a rampup margin [47,48]. When curtailing, some electricity is always lost; however,
this is acceptable if the small losses enable larger VRE installations.
3.2.2
Energy storage
Storage of energy is a way to time-shift the energy supply, which makes temporal mismatches of supply and demand tolerable. Recently, interest in energy
20
Energy system flexibility (Publication I)
storage has increased because of large-scale integration of renewable energy,
which has led to advances in storage technologies [49]. Well-known examples
of energy storage are hot water tanks (for storing heat) and fuel containers (for
fuels), which are economical and technologically simple. Storing electricity is,
however, more difficult due to high costs and technological challenges.
Options for storage of electricity with different characteristics are numerous.
Publication I considers pumped hydro power energy storage (PHES), compressed air (CAES), flywheels, batteries, hydrogen, superconducting magnets,
and supercapacitors. Some technologies are more suitable for providing shortterm power, whereas others are more suitable for long-term storage. The chosen
technology depends on the case, as none of the available systems can provide all
of the desired features: low cost, high efficiency, long lifetime, and high power
density [50,51].
PHES accounts for 99% of the total storage capacity in the world and is thus
by far the most utilized technology, but CAES and lead-acid batteries are also
considered as mature technologies [52]. In small appliances, lithium-ion batteries are currently the most widely used technology, due to their high efficiency,
good energy density and low self-discharge rate. For large-scale grid services,
however, they are still overly expensive [53,54].
The energy efficiency of the power system could be increased with energy storage, as base-load power plants would be able to continue their high-efficiency
production even during periods of low demand. Also, the economic efficiency
would be enhanced by using stored energy instead of peak-load power plants
with high marginal costs [55].
3.2.3
Power-to-X
Power-to-X (P2X) refers to technologies through where electricity is converted
to other energy forms, such as heating or cooling, gas (methane or hydrogen),
and power for electric vehicles. This increases the flexibility of the system by
providing new options (i.e., degrees of freedom) in the end use of energy. The
electricity for P2X can be taken from surplus VRE production or from power
markets when the price is low or even negative. These technologies make it possible to reduce or even avoid curtailment of renewable energy.
In power-to-heat (P2H) or, more generally, power-to-thermal (P2T), electricity is converted into thermal energy (i.e., heating or cooling). The main technologies for this purpose are heat pumps and electric boilers connected to thermal
storage or air-conditioning units. Typically, thermal energy dominates in final
energy use, and it is easier and more economical to store. Advantages of P2H
and P2T also include reductions in emissions of heat production from fossil
fuels [56].
Excess electricity can also be used in power-to-gas conversion (P2G) – that is,
production of high-value gas products, such as hydrogen or methane. Hydrogen
can be produced by electrolysis or photoelectrolysis. Synthetic methane is typically produced through hydrogen and the Sabatier reaction [57,58].
21
Energy system flexibility (Publication I)
Electric vehicles (including plug-in electric vehicles) provide a way to store
energy. Combining the batteries of the vehicles actively with the grid is called
vehicle-to-grid strategy (V2G). The vehicles stand idle most of the time, with
relatively long (8–12 h) charging windows but short (90 min) charging times,
resulting in considerably flexible scheduling of storage [59,60].
3.2.4
Demand-side management
Demand-side management covers a wide range of technologies that affect the
end-use electricity consumption. Contrary to the traditional way of controlling
the energy supply, its purpose is to modify the demand to better match the supply. It can be roughly divided into three categories: reducing, increasing, or rescheduling the demand. All of these can be used to even out the demand by clipping high peaks and filling valleys of low consumption [61]. Load shifting (i.e.,
rescheduling) typically requires intermediate storage. Examples of demandside management include controlling the electric heating in buildings and postponing the use of washing machines. An industry with high power consumption
may reduce or postpone its production when power consumption is high and
electricity is expensive.
DSM is able to offer several benefits to power systems and markets. It balances
the demand and supply in terms of both energy and power, and it operates in
various time scales. Most DSM applications are used in the time scale of 1–12 h,
where the largest fluctuations of VRE supply also occur [62]. DSM smooths the
operation of power markets by reducing price peaks and the average spot price
[63]. DSM can also lead to economic savings by removing or postponing the
need for infrastructure expansions, by reducing transmission and distribution
losses, and by facilitating energy savings [64–66].
3.2.5
Novel grid infrastructure
The flexibility of power systems requires sufficient transmission capacity [67].
Spatial differences in power supply and demand can be balanced by a robust
and well-designed grid and appropriate grid codes [48]. Strong transmission
lines are also essential for energy markets to function properly. Supergrids,
smart grids, and microgrids are proposed novel technologies to help with the
flexibility challenge.
Supergrids refer to strong transmission networks that typically use high-voltage direct current (HVDC) in energy transfer. HVDC technology enables smaller
electric losses and could be more economical in long-distance power transfer.
Thus, it could be used to connect large remote VRE installations with demand
[68,69].
Recently, smart grids have generally been considered as a key technology for
efficient VRE integration. The goal of smart grids is to optimize the operation of
the energy system and to enable new and more efficient energy market structures by connecting energy producers, consumers, network companies and storage facilities intelligently together. This requires automation, robust two-way
communication technology and advanced metering. Smart grid development
22
Energy system flexibility (Publication I)
aims at achieving improved power supply reliability and reduced ecological impact, while avoiding costly investments by finding new options for securing energy supply [21,22].
Microgrids are small local energy networks that can meet the demand of local
consumers. These may include local generation (micro-CHP, VRE), storage, and
demand-side management. They can be built in isolated locations, but also as a
part of a larger grid system, where the microgrid significantly improves the local
power reliability by offering a solution if the power supply from the larger system fails. Furthermore, microgrids could be used as a component to balance
voltage fluctuations in larger systems [70,71].
23
4. Spatiotemporal consumption profiles
(Publication II)
To better understand the operation of energy systems of cities, one must have
detailed information about consumption in terms of both space and time. Detailed spatiotemporal data about consumption allows for better planning of
large-scale renewable energy generation and smart-grid operations, as well as
understanding their effects on energy networks and the overall system operation.
Typically, energy consumption in cities is modeled either temporally or spatially. Publication II presents a model to combine these dimensions and to create so-called spatiotemporal load profiles. It is based on the spatial division of
temporal profiles P(t) to P(x,y,t). The model guarantees that the original load
profile P(t) stays unmodified. For spatial division, the model uses spatial data
of urban structure, such as population and work-place distribution, but also
building type and land-use data.
The accuracy of the profiles produced depends on the available input data. The
model can also be used with coarse data by utilizing different urban structure
models, as demonstrated in Publication II with the Shanghai case.
4.1
Input data
The model is based on classifying energy consumers into consumer classes. A
typical rough division of consumer classes involves using classes: residential,
services and manufacturing. These can be further divided into sub-classes: the
‘residential’ class may include single-family houses and apartment buildings,
whereas the ‘services’ class may include offices, healthcare, educational buildings, restaurants and commercial buildings. This division depends on the data
available in the city.
As input data, the model requires the consumption time series of the whole
city, and the consumer classes’ annual consumption numbers and temporal
base profiles. In the absence of annual consumption numbers, floor areas (m2)
can also be used with typical energy intensities (MWh/m2/year) of the classes.
The base profiles are relative load profiles revealing how the consumption of a
consumer class is distributed over time. Their yearly integral over time (or sum
over discrete time steps) is equal to 1.
25
Spatiotemporal consumption profiles (Publication II)
For spatiality, the data required are data about the spatial distribution of the
classes in different districts of the city. These data sets include, for example,
population densities and floor area data in the districts. If detailed spatial data
about the urban structure and the building locations are not available, mathematical models for urban structure can be used. For example, in Publication II,
the following exponential density model [72] was used:
మ
݀ሺ‫ݔ‬ሻ ൌ ݀଴ ݁ ିఈሺ௫ି௫బ ሻ ൅ ‫ܭ‬
(1)
where ݀଴ describes the maximum density of an urban structure, ‫ ݔ‬is the distance from the city center, ‫ݔ‬଴ is the location of the maximum density, ߙ is the
width parameter of the density distribution, and ‫ ܭ‬is a constant describing the
minimum density throughout the area.
4.2
Formation of spatiotemporal load profiles
A visual illustration of the model is presented in Figure 1. For clarity, it shows a
case with no sub-classification of consumer classes. If sub-classes are used, the
process for them is similar to the one described here, with the total city load
profile replaced by a load profile of the original consumer class.
4.2.1
Scaling base profiles to match the city demand
The procedure begins by multiplying the base profiles ݂௜ ሺ‫ݐ‬ሻ with the corresponding annual consumption ܳ௜ . Subscript ݅ refers to consumer classes. The multiplification forms a bottom-up type of preliminary load profiles for the classes.
Their sum represents the total consumption of the city.
If the original total city profile ܲୡ୧୲୷ ሺ‫ݐ‬ሻ is known, the summed profile (probably) does not match it. Therefore, independently at each time step, the summed
profile is scaled to match the city profile ܲୡ୧୲୷ ሺ‫ݐ‬ሻ. The scaling factor is indicated
by ߣሺ‫ݐ‬ሻ. The final load profiles ܲ௜ ሺ‫ݐ‬ሻ for the consumer classes are, thus, of the
form
ܲ௜ ሺ‫ݐ‬ሻ ൌ ߣሺ‫ݐ‬ሻܳ௜ ݂௜ ሺ‫݅ݐ‬ሻ
(2)
where
ߣሺ‫ݐ‬ሻ ൌ
4.2.2
ܲୡ୧୲୷ ሺ‫ݐ‬ሻ
σ௜ ܳ௜ ݂௜ ሺ‫ݐ‬ሻ
(3)
Sub-class division
If sub-class division is used for consumer classes, profiles ܲ௜ ሺ‫ݐ‬ሻ must be further
divided into ܲ௜௝ ሺ‫ݐ‬ሻ, with ݆ indicating the sub-class. This is done similarly to
forming class profiles ܲ௜ ሺ‫ݐ‬ሻ. The sub-class base profiles ݂௝ ሺ‫ݐ‬ሻ are multiplied with
annual consumption of the sub-class; that is, ܳ௝ or ‫ܣ‬௝୲୭୲ ‫ܫ‬௝ , where ‫ܣ‬௝୲୭୲ is the total
floor area of sub-class ݆ in the city, and ‫ܫ‬௝ is its typical energy intensity.
These multiplied base profiles are again scaled up or down so that their sum
profile corresponds to the load profile of the upper class ݅. That is
26
Spatiotemporal consumption profiles (Publication II)
ܲ௜௝ ሺ‫ݐ‬ሻ ൌ ߣ௜ ሺ‫ݐ‬ሻܳ௝ ݂௝ ሺ‫ݐ‬ሻ
(4)
where
ߣ௜ ሺ‫ݐ‬ሻ ൌ
ܲ௜ ሺ‫ݐ‬ሻ
σ௝ ܳ௝ ݂௝ ሺ‫ݐ‬ሻ
(5)
Figure 1. Schematic representation of the model for creating spatiotemporal load profiles. Reproduced with permission from Publication II. Copyright 2014 Elsevier.
27
Spatiotemporal consumption profiles (Publication II)
4.2.3
Spatial division
After temporal load profiles for consumer classes and sub-classes are formed,
they are divided spatially according to any spatial data available. Here it is assumed that the division is made by floor area distribution ‫ܣ‬௝ ሺ‫ݔ‬ǡ ‫ݕ‬ሻ, but any other
density unit could be used (in Figure 1, general density݀௝ is used). The load profile of class ݅and sub-class ݆ for location ሺ‫ݔ‬ǡ ‫ݕ‬ሻ is
ܲ௜௝ ሺ‫ݔ‬ǡ ‫ݕ‬ǡ ‫ݐ‬ሻ ൌ
‫ܣ‬௝ ሺ‫ݔ‬ǡ ‫ݕ‬ሻ
ܲ௜௝ ሺ‫ݐ‬ሻ
‫ܣ‬௝୲୭୲
(6)
The final total load profile of location ሺ‫ݔ‬ǡ ‫ݕ‬ሻ is formed by summing these classby-class load profiles of the location together:
ܲሺ‫ݔ‬ǡ ‫ݕ‬ǡ ‫ݐ‬ሻ ൌ ෍ ෍ ܲ௜௝ ሺ‫ݔ‬ǡ ‫ݕ‬ǡ ‫ݐ‬ሻ
௜
28
௝
(7)
5. Multi-carrier energy networks (Publication III)
The energy system consists of several energy production, conversion, and distribution processes. Typical examples are conversion chains from primary energy into a utilizable form – in other words, energy commodities (or energy carriers), such as fuels, electricity and heat. These are transported and distributed
to consumers by energy networks, such as power grids, gas distribution lines,
and district heating networks. The commodities can also be converted into each
other (e.g., gas into electricity and heat, or electricity into heat and cold). Note
that low-exergy heat cannot be converted into high-value energy products.
A multi-energy concept refers to technologies and practices in which different
energy carriers are distributed, converted between each other, and stored optimally for the whole system. For example, the share of thermal energy in the final
energy mix of built environments could be two to three times higher than that
of electricity [12]. Thus, taking the thermal energy system into account when
operating the electric system could provide significantly more flexibility and options for matching demand and supply. Storage of thermal energy is also significantly easier and cheaper. Another example is that of power-to-gas technologies, which convert renewable electricity into gas when excess production is
available. Multi-energy thinking makes it possible to oversize the VRE production by utilizing the electricity that could not be fed into the electric distribution
system, thus yielding notable VRE production during modest wind or solar conditions.
Publication III introduces a concept for multi-energy simulations based on
spatiotemporal data on urban energy consumption. In Publications IV and VI,
the model is used for energy system analysis with more detailed input data.
5.1
Structure of the multi-carrier energy systems model
The energy system is modeled here with a combination of nodes that are connected by lines forming networks, though not all energy networks will necessarily have the same spatial coverage. Depending on the case and the desired
spatial accuracy, a single node can represent a neighborhood, a block or a single
building of a city. Each node is typically equipped with time series of energy
29
Multi-carrier energy networks (Publication III)
consumption (such as electricity ܳ௘ , heat ܳ௛ , and cold ܳ௖ ). Furthermore, a generic node may contain different functionalities, such as energy production (ܲ),
storage (‫)ݏ‬, and mutual conversion (ܿ) of energy carriers (electricity, fuel, heat,
cold). In practice, the nodes may not include all the functionalities simultaneously but may contain some simple combination of them. Figure 2 shows an
example of a generic node. Since the nodes are connected to each other by networks, each node also has incoming and outgoing energy flows (݂୧୬ and ݂୭୳୲ ).
Figure 2. Representation of a generic node with different functionalities: energy flows (݂), production (ܲ), consumption (ܳ), storage (‫)ݏ‬, and mutual conversion of energy forms (ܿ). Subindex ݁ refers to electricity, ݄ to heat, ܿ to cold, ݂ to fuel, and ߢ to any of these.
Figure 3 shows an example of two energy networks covering the area divided
into nodes. The networks are connected to each other by intersecting nodes, in
which the conversion occurs and the energy moves between the networks. These
intersections are marked with dashed purple lines in the figure. The energy networks (power grid, district heating (DH) and cooling networks, and gas pipelines) may be connected to a higher-level transmission system. These nodes are
called energy connectors and marked with circles in the figure. The connector,
in a power grid, for example, is a transformer station, while in a district heating
network it is a heat exchanger or a DH substation.
Figure 3. Two energy networks (power grid and district heating network) connected to each other
by intersections (purple dashed lines). Circles represent energy connector nodes (blue:
transformer station; red: district heating station with heat exchangers).
30
Multi-carrier energy networks (Publication III)
5.2
Energy balance in nodes and flows in network
௧
Based on the conservation of energy, the energy balance ‫ܤ‬௞ǡ௡
(i.e., match of sup-
ply and demand) is calculated for each node ݊ and each energy form ݇ at time
point ‫ݐ‬
௧
௧
௧
௧
௧ା
௧ି
௧
‫ܤ‬௞ǡ௡
ൌ ܲ௞ǡ௡
െ ܳ௞ǡ௡
൅ ‫ݏ‬௞ǡ௡
൅ ܿ௞ǡ௡
െ ܿ௞ǡ௡
െ ‫ݖ‬௞ǡ௡
௧ା
ܿ௞ǡ௡
(8)
௧ି
ܿ௞ǡ௡
where the terms
and
represent the conversion of other energy forms
௧
into form ݇ and vice versa, and the term ‫ݖ‬௞ǡ௡
is the curtailed energy production.
௧
The storage term ‫ݏ‬௞ǡ௡ can have either a positive or a negative value depending
on whether the storage unit is discharging (+) or charging (−). It is assumed that
the energy losses are included in the corresponding terms.
Conversion terms may be defined with the efficiency of the conversion process
between energy forms
௧ା
௧
ܿ௞ǡ௡
ൌ ෍ ߟ఑՜௞ ‫ݍ‬఑ǡ௡
(9)
఑
఑ஷ௞
௧
where ‫ݍ‬఑ǡ௡
represents the amount of the other energy forms allocated for conversion into energy form ݇. The index ߢ covers all other energy types except ݇.
In the opposite case, the energy form ݇ is converted into other forms
௧ି
௧
ܿ௞ǡ௡
ൌ ෍ ߟ௞՜఑ ‫ݍ‬௞ǡ௡
(10)
఑
఑ஷ௞
The efficiency is typically between 0 and 100%, but can also have values greater
than 1, such as when using heat pumps in power-to-heat conversion, where ߟ
corresponds to coefficient of performance (COP).
Eq. (8) states how much of the energy consumption of the node can be satisfied by local production, discharging the local storage, and converting other energy carriers. On the other hand, the equation reveals if there is surplus production in the node. If the balance differs from zero, the energy flow in the network
must compensate for that:
௧
௧
௧
‫ܤ‬௞ǡ௡
ൌ ݂୧୬ǡ௞ǡ௡
െ ݂୭୳୲ǡ௞ǡ௡
(11)
Flow terms heavily depend on the structure of the network and the operating
principles of the whole energy system. With a simple non-looped network, the
power flows in the network can be determined from the condition that a node
must deliver all the power to the nodes downstream of it. For a line ݈ in the network, this yields a condition
௧
௧
݂௞ǡ௟
ൌ ෍ ‫ܤ‬௞ǡ௡
(12)
௡‫א‬ୢୱሺ௟ሻ
where notation ݊ ‫•† א‬ሺ݈ሻ refers to nodes that are downstream of line ݈. “Downstream” refers to the nodes that are network-wise further from the energy connector (e.g., transformer in a power grid). See Figure 4 for an explanation. For
a looped network structure, more complicated iterative calculation methods are
required, such as the Newton–Raphson method [73].
31
Multi-carrier energy networks (Publication III)
Figure 4. Simple network to explain the terms upstream and downstream. The node marked with
a white circle is the energy connector (transformer, heat exchanger). The nodes with the red
background are upstream of the node with the yellow star, while the nodes with the green
background are downstream of it.
5.3
Energy system operation in simulations
To run the energy system simulations, it is necessary to predetermine what to
do in cases where the local energy production does not match the consumption,
or in the event that the carrying capacity of the network limits the flows. Which
actions are taken first? For this purpose, different strategies may be utilized depending on the case. A few examples are listed below:
1) No conversion–no storage:
All excess energy production in the nodes is fed to the networks; if they get congested, the VRE production is curtailed.
2) Local use with conversion and storage:
The target is to consume all locally produced energy in the same node. Energy
is converted from one form to another to meet the demand of all energy forms.
Nodes use their own storage units to avoid feeding energy into networks.
3) City-wide use:
The purpose is to keep all produced energy within the city. Conversion may take
place in several nodes, and it is applied when needed due to congested networks
or a mismatch between supply and demand. Some energy forms may be stored
in the nodes.
4) National transmission:
Excess electricity or fuel is fed to national networks through energy connector
nodes. Conversion is applied only if the network gets congested, preventing the
use of national transmission. Heat is produced within the city or locally in the
nodes.
Typically, the energy balances are calculated first, and the calculation of power
flows (as in Eq. (12)) requires balances as input data. Some of the above-mentioned strategies, however, require that the possible grid congestion is checked
32
Multi-carrier energy networks (Publication III)
when determining the actions in the nodes and affecting the energy balance. In
these situations, iterative calculations are needed, starting from the end points
of the network (i.e., the nodes with no downstream nodes) and proceeding
node-by-node towards the energy connector node. The energy balance is first
calculated without any further actions. Next, the power flow in the upstream
link is checked; in the case of congestion, the required amounts of flexibility
methods such as conversion or storage are utilized. The power flow is then recalculated with an updated value of energy balance.
33
6. Optimal use of an energy system under large-scale VRE (Publication V)
Energy production is gradually transforming from centralized, power plantbased generation into variable and distributed renewable generation. However,
the old and new systems will still have to operate side-by-side for future decades. Thus, it is highly important to understand the effects of increasing variable
renewable generation on the existing system, especially as the current power
plant mix may not be designed for large-scale VRE.
Publication V introduces a mixed-integer linear optimization model to find the
most economical ways to operate the whole energy system (both electricity and
heating) under large-scale VRE, consisting of traditional energy production
plants (power plants, CHP plants, heat boilers), different flexibility technologies, energy spot markets, and emission costs.
The model minimizes operating and fuel costs and maximizes revenues from
energy sales, while the demand must always be satisfied. It can be used to model
energy systems from single buildings to the city scale. Investment costs of new
technologies are not included, as the focus is on the effect of VRE integration on
existing power plants’ operation and economy, and thus on the operation of the
whole system.
6.1
Model principles
The model is based on mixed-integer linear programming (MILP). This formulation guarantees short calculation times and simple modifying of the model by
easily adding or removing decision variables and boundary conditions.
The model runs time step by time step, seeking the most economical solution
that fulfils the requirement of always meeting the energy demand. The optimal
solution is determined by considering several technical limitations in energy
balance, energy production plants, and flexibility technologies. Figure 5 shows
schematically how both internal and external factors affect the system operation
and thus its optimization.
35
Optimal use of an energy system under large-scale VRE (Publication V)
Figure 5. External and internal conditions affect the optimization problem. Reprinted with permission from Publication V. Copyright 2016 Elsevier.
Figure 6 presents a schematic representation of the model principles. Time
series of VRE production (e.g., wind and solar power) and energy consumption
form the basis of the model (Point #1 in the figure). As input, properties of the
existing energy system are also given (Point #2), including the technical limitations of energy production plants and energy networks, as well as storage and
flexibility technologies.
The mixed-integer linear optimization problem (Point #3) is solved at each
time step for a given number of future time steps (i.e., optimization horizon). In
forecasting the future consumption and VRE production, forecast errors should
be considered, and thus this option is included in the model as well.
As output (Point #4), the model produces optimal time series of power plant
operation and flexibility technologies, as well as the costs and profits from satisfying the energy demand. Even though the optimization algorithm yields optimal time series for the whole optimization horizon at each time step, only the
values of the first time step of the horizon are used in the output profiles, since
the rest of the time steps are updated during the optimization of the following
time steps.
6.2
6.2.1
Input and output
VRE production, consumption, and technical parameters
Energy consumption (electricity and heat) and VRE production time series are
given as input. They are used in boundary conditions of the optimization problem to guarantee that the energy supply always matches the demand.
The forecast errors can be included in the input time series by modifying the
actual time series by error terms, which could be formed based on Wiener processes (i.e., random walks) [74]. The idea in doing so is to form separate Wiener
processes starting from each time step. These processes are then cumulatively
summed up so that the error for the next time step (‫ ݐ‬൅ ͳ) is based on the Wiener
process of that time step. The errors of additional time steps (‫ ݐ‬൅ ʹǡ ‫ ݐ‬൅ ͵ǡ ǥ) are
based on the sum of Wiener processes of all time steps between the current and
the forecasted time step. Finally, the forecast errors are scaled so that their expected values match the measured forecast errors.
36
Optimal use of an energy system under large-scale VRE (Publication V)
Figure 6. Schematic illustration of the MILP model to find cost-optimal ways to operate energy
systems under large-scale VRE. Reprinted with permission from Publication V. Copyright
2016 Elsevier.
The technical parameters of power plants and energy systems are incorporated as boundary conditions of the optimization problem as well. These include
characteristics of power plants, such as maximum and minimum power outputs,
ramping rates, and start-up and shut-down times. For CHP plants, it is necessary to define how tightly power and heat outputs are connected to each other.
Flexibility measures are defined as well: size of the heat storage, power transmission capacity to upper-level transmission system (e.g., national grid), coefficient of performance (COP) of P2H conversion, and upper limits for curtailment and P2H.
37
Optimal use of an energy system under large-scale VRE (Publication V)
6.2.2
Economic parameters
Economic parameters are used as coefficients in the objective function of the
optimization, since the target is to minimize the costs and maximize the revenues. The cost of producing energy in power plants is a combination of several
factors, including the fuel price, operations and maintenance (O&M) costs,
emission prices, and taxes. In this work, the production costs in CHP plants are
divided into electricity and heat in relation to the nominal outputs of the energy
forms.
Ramping the power plant, especially if it is designed for base-load operation,
has a degrading effect on the plant, and thus it is thought to have a cost. Thus,
we have included a cost for ramping in the model, which is defined as euros per
mega-watt change in the output.
Income comes from selling energy to consumers within the grid area of the
energy company. For district heating sales, seasonal pricing is commonly used,
meaning that the price changes a few times per year. The selling price for electricity typically depends on the spot price of electricity, which is used in this
work as the selling price. The price is different when electricity is sold to power
markets and when it is bought from them. In this work, the spot price is used,
but it is modified with a margin to take the transfer prices into account. The spot
price is considered an external factor (Figure 5), which requires that the investigated area must be small enough compared to the whole energy market of
which it is a part, so that the internal phenomena do not have significant effect
on the market prices (e.g., the consumption of Helsinki city (Finland) is only
about 1% of the whole Nordic–Baltic spot market).
Operating the flexibility measures can also have costs. Thus, it is possible to
define the costs of P2H conversion, curtailment, and storage of heat. The model
also allows for the heat to be wasted, but typically the cost for this is set so high
that it is avoided as long as possible.
6.2.3
Output
During the simulations, the model produces several output time series. These
include the power outputs from the energy production plants, the amounts of
electricity and heat sold to consumers, and electricity sold to or bought from the
spot markets. Additionally, the flexibility measures are recorded during the simulation: P2H conversion, the state of the heat storage, curtailed VRE production
and wasted heat.
All these actions together form the total costs and revenues from running the
energy system. From those time series, additional results can be calculated, such
as yearly energy balances, emissions, and capacity factors.
38
Optimal use of an energy system under large-scale VRE (Publication V)
6.3
6.3.1
Optimization of energy system operation
Objective function
During the simulations, the optimization is run at each time step considering
the current time step and a number of future time steps (i.e., optimization horizon; see Figure 6). The complete mathematical formulation is presented in Publication V. The linear optimization problem follows the formulation
‹ ෍ሺ’”‘†—…–‹‘…‘•–• ൅ „ƒŽƒ…‹‰…‘•–•
௧
൅ •–‘”ƒ‰‡…‘•–• െ ”‡˜‡—‡•ˆ”‘•ƒŽ‡•ሻ
subject to
(13)
a) Technical properties of energy production plants
b) Technical properties of balancing methods
c) Balance of energy demand and supply
The costs from all time steps in the optimization horizon are summed up, and
the revenues are subtracted from them to form the objective function. Production costs include fuel, O&M, and emission costs of energy production, as well
as possible taxes. Furthermore, ramping costs from changing the output power
of the plants are included here. Auxiliary variables are needed in the formulation for ramping to keep the model linear. Because the ramping may be positive
or negative, absolute values would be needed in the objective function to keep
the cost positive, which would break the required linearity. Balancing costs
cover the O&M costs of the technologies, but the lost profits from non-sold electricity are excluded. The revenues from sales include the income from the energy
sold to consumers in the studied area. The effect of the electricity that is produced in the city but used in balancing technologies instead of sold it, is shown
as decreased sales in the revenue term of the function. O&M costs of the energy
storage are included in the storage costs.
6.3.2
Operation principles of an energy system – boundary conditions
The energy production plants typically have minimum and maximum values between which the output varies when the plant is on. In addition, start-up and
shut-down procedures take time. These characteristics are included in the
model by binary variables defining the plants’ current operational state (on,
start-up, off, or shut-down), and, depending on the state, corresponding limits
are set for output power. Separate boundary conditions are defined to guarantee
that the operational state does not change too quickly (i.e., the start-up and
shut-down times are obeyed).
The plant operation is additionally restricted by ramping limits, or the maximum allowed change in output power between time steps. Some plants (e.g.,
heat boilers designed for peak-load operation) may have less strict conditions,
such as only limiting the maximum output but allowing quick ramping.
The boundary conditions for balancing methods are kept as simple as possible.
Thus, their use is limited between zero and a maximum value. For curtailment,
39
Optimal use of an energy system under large-scale VRE (Publication V)
the upper limit could be defined as an absolute limit or a relative amount of the
momentary production.
An important boundary condition is the balance between demand and supply.
In this condition, one must include all possible options for energy. The production must be equal to the consumption, taking into account curtailment, P2H
conversion, charging and discharging the storage units, and the electricity import from/export to the national transmission system. If the heat supply exceeds
the demand after all the balancing measures are fully utilized, the extra heat
needs to be wasted (but with high cost).
40
7. Results for Helsinki
Helsinki, Finland, is a northern city in a global perspective with around
600,000 inhabitants. Its final energy use is dominated by heat, mainly due to
its location (60 °N). Its current energy supply is almost completely dependent
on fossil fuels, though CHP production and extensive district heating result in
high fuel efficiency. The city, however, has a target of becoming carbon-neutral
by 2050 [75], which requires drastic redesign of the energy system. For these
reasons, the city was chosen as the main case study for the thesis.
7.1
Current energy system in Helsinki
The current energy sources of the municipal energy company of Helsinki are
shown in Figure 7. The energy supply of the city is based on three large CHP
plants using natural gas or coal as fuel, providing both electricity and heat to
consumers. These plants’ production capacity is around 1,000 MW of electricity
and 1,300 MW of heat. For cold peak conditions during the winter, there are ten
heat-only boilers in the city, with a total capacity of more than 2,000 MW. Additionally, there is a heat pump station able to produce 90 MW of heat to the
district heating network. These energy generation plants are marked on the map
in Figure 8.
Annually, around 4 TWh of electricity is consumed in the city [76]. However,
the total production capacity of the municipal energy company is 7 TWh, but all
nuclear and almost all of the renewable production capacity (Figure 7 left) are
located far from the city. The power plants in Helsinki typically produce somewhat more electricity than the city consumes. For example, in 2014, the CHP
plants produced 4.7 TWh.
The annual heat consumption in the city varies between 6–7.5 TWh depending
on the weather [76]. Over 90% of the heat load is covered by a district heating
system that has 1,300 kilometers of hot water pipelines [77,78]. In 2015, around
6.4 TWh of district heat was supplied in Helsinki, of which 88% was covered by
CHP, 7% by heat pumps, and the rest mainly with heat boilers [79].
41
Results for Helsinki
1%
13
%
20
%
Natural gas
49
%
18
%
7%
Coal
40 %
52 %
Oil
Nuclear
Renewables
Figure 7. Left: The energy sources of the electricity production of the municipal energy company
in Helsinki in 2015, including the shares in the production capacity outside the city (nuclear
+ renewables). Right: The energy sources of the district heating in Helsinki in 2015. [76]
Figure 8. Map of Helsinki city, with total floor area distribution. Large CHP plants are marked with
A–C (A natural gas, B & C coal), heat-only boilers with D–F (D natural gas, E oil, F coal), and
heat pump station with G. Reproduced with permission from Publication V. Copyright 2016
Elsevier.
Examples of load profiles of electricity and district heating are shown in Figure
9 and Figure 10, respectively. The heat load includes both space heating and hot
water. The space heating load depends almost completely on the outside temperature, whereas the hot water demand is fairly constant throughout the year.
Due to space heating, the overall heat load is strongly concentrated in the winters only, as the consumption during the summers is mainly hot water consumption. In contrast to heating, the electricity consumption depends more on the
actions of people than on weather conditions, even though weather also has a
clear effect on the load profile. The electricity consumption is, however, distributed more evenly on a yearly scale, but there are clear weekly and daily cycles.
42
Results for Helsinki
Figure 9. Temporal load profile of electricity consumption in Helsinki during one year. Reproduced with permission from Publication II. Copyright 2014 Elsevier.
Figure 10. Temporal load profile of heat consumption in Helsinki during one year. Reproduced
with permission from the Supplementary material of Publication V. Copyright 2016 Elsevier.
7.2
7.2.1
Spatiotemporal consumption profiles (Publication II)
Input data
The official district division (before 2013) of Helsinki city was used to divide the
temporal load profiles (Figure 9 and Figure 10) spatially. Some slight modifications have been made since then to the district borders. The districts are shown
in Figure 8. The consumers were divided into three main classes: households,
services, and industry. These were further divided into sub-classes, which are
listed in Table 2. The subclass division is the same one that is used in the official
statistics of the city [80].
So-called standard load profiles based on German experiences [81] are used
as the base profiles ݂௜ and ݂௝ required in Eqs. (2) and (4). The load profiles and
their use with the consumer classes are explained in Publication II. The total
43
Results for Helsinki
annual consumptions of the consumer classes and subclasses ܳ௜ and ܳ௝ are
from the Helsinki Environment Centre [82]. In this work, values from 2011 were
used; in that year in Helsinki, the household electricity consumption was 1.3
TWh, service sector consumption was 2.9 TWh, and industry consumption was
0.3 TWh. In 2015, the residential consumption increased by 10%, but service
and industry consumption decreased by 8–11% compared to the 2011 values.
The building distributions of different consumer types (i.e., subclasses) were
obtained from the statistics of the city [83]. The total floor area distribution is
shown in Figure 8. Two examples of building distributions of consumer classes
are shown in Figure 11: a) households and b) office buildings. The residential
buildings are located more evenly throughout the city, whereas the office buildings are mainly concentrated in the downtown area.
Table 2. Consumer class division used in Helsinki. Reproduced from Publication II with permission. Copyright 2014 Elsevier.
Households
Services
Industry
Detached houses
Commercial
Industry
Row houses
Office
Storage
Apartment houses (less than 4 floors)
Traffic
Apartment houses (4 floors or more)
Healthcare
Recreational
Education
Figure 11. Examples of floor area distributions of consumer classes in Helsinki; a) households,
b) office buildings.
7.2.2
Spatiotemporal consumption profiles – results
Figure 12 illustrates how the electricity consumption behaves spatially during
a typical working day in Helsinki. In the city center and other office areas, the
demand peaks at noon or in the late morning hours, whereas in the residential
areas further from the city center, the consumption peaks around 7–8 pm. A
milder peak occurs during the morning hours of 6–8 am, before people leave for
work.
The highest peaks in electricity load in the neighborhoods over the whole year
are found in the city center in early February. A level of 80 MW/km2 can be
reached during the coldest times. In residential areas some kilometers away
from the center, the peak load reaches only 3–6 MW/km2.
44
Results for Helsinki
Figure 12. Spatial electricity load of an average working day in Helsinki between a) 2–3 am, b)
11–12 am, c) 7–8 pm. Reproduced with permission from Publication II. Copyright 2014 Elsevier.
45
Results for Helsinki
Figure 13. Relative load profiles from three districts of Helsinki. Kluuvi is a district in the city center
with a high share of office buildings; Punavuori has both apartment buildings and offices; and
Puistola is an almost completely residential area with many detached houses. Reproduced
with permission from Publication II. Copyright 2014 Elsevier.
Relative load profiles of three selected districts are illustrated in Figure 13 (locations shown in Figure 12). Kluuvi is an office and commercial district in the
city center, whereas Puistola is a neighborhood of mainly detached and row
houses in the northern part of the city. Finally, Punavuori (in the immediate
vicinity of the city center) is a mixture of these two, with a remarkable share of
apartment buildings but also many offices and service buildings. In the downtown areas (Kluuvi), electricity consumption peaks during the working hours,
whereas in the residential area (Puistola), the peaks occur in the morning and
in the evening. The peak power consumption in these districts is more than two
times higher than the off-peak consumption during the night hours. The profiles
are also significantly different on the weekends in the different types of areas.
The residential areas have high consumption numbers during weekends,
whereas in the office areas, the consumption stays lower on weekends than
weekdays.
Validation of the results was difficult due to the lack of data from the districts.
Therefore, a simple comparison was made with aggregated annual values. The
consumption of the residential sector was 1,500 GWh according to the model,
whereas the measured value was 1,300 GWh. The model showed a consumption
of 2,800 GWh (with 2,900 GWh being the measured value) for the service sector, while it showed 260 GWh (290 GWh measured value) for the industrial sector. The modeled values can be considered satisfactory, with a maximum deviation of 15% in the results.
7.3
P2H conversion enables more VRE (Publications III and IV)
As the city of Helsinki has a well-established district heating network that accounts for more than 90% of the city’s heat consumption, it was studied in this
work whether the flexibility of the energy system could be increased through
power-to-heat conversion, which could lead to larger VRE capacity. Converting
46
Results for Helsinki
electricity into heat is associated with poor exergy efficiency (second law of thermodynamics), even though the first-law efficiency (energy quantity) could be as
high as 300% when using heat pumps.
There are, however, several factors in favor of P2H. Firstly, the heat demand
in a built environment is large [12]. Secondly, the heat system has greater inherent inertia compared to the electric one, and the storage of heat is easier and
cheaper. Furthermore, by using heat pumps, the thermal output could be increased with the same amount of electricity. Finally, P2H enables the use of
more VRE in the electricity consumption by allowing the VRE production to be
oversized, and as the capacity factor of VRE is clearly less than 1, the amount of
converted energy in P2H is not proportional to the curtailment level of power
but closer to the capacity factor of VRE (0.2–0.35 for wind) [56,84] (Publication
IV).
7.3.1
Wind power with P2H and thermal storage
Integration of large-scale wind power was spatiotemporally analyzed in detail,
together with a large P2H scheme. The consumption values of 4.4 TWh for electricity and 6.6 TWh for heat were used in the analysis, and the spatial consumption profiles introduced in the previous chapter (Chapter 7.2) were used. The
wind power scheme consists of three wind power farms that are connected to
the grid close to the CHP plants (locations shown in Figure 8). The connection
points were chosen because of the strong electric network in the locations. The
amount of wind power was varied from zero to up to 4,000 MW with a unit
output of 3.14 GWh/MW. P2H conversion was done with electric resistance
(COP = 1) or heat pumps (COP = 3), and three different heat storage schemes
were included (no storage, limited to 100 GWh, and unlimited).
The self-use limit of VRE is the maximum amount of installed capacity where
the energy production ܲ does not exceed the consumption ܳ on an hourly basis
( ܲ always ൑ ܳ ). With wind power, the self-use limit in Helsinki is around
400 MW, which covers 27% of the city’s electricity consumption. If the wind
power installations are increased above this limit, part of the production needs
to be curtailed or exported outside the city.
Figure 14 shows the share of wind power in total electricity consumption and
the share of P2H-produced heat in total heat demand with different amounts of
installed wind power and different P2H strategies. The linear increase of the
share of wind power in electricity consumption starts to slow down after the
self-use limit (400 MW). Oversizing the wind power capacity, however, provides
benefits on the electricity side as well, as the share of wind in electricity consumption continues to grow after the self-use limit. The surplus production is
converted into heat. With 2,000 MW of wind power and 100 GWh of heat storage, more than 60% of the electricity and 70% of the heat that is needed could
be produced with wind power. By comparison, without thermal storage, only
slightly more than 30% of the heat could be delivered and more than half of the
wind-based heat would be lost.
47
Results for Helsinki
% of electricity
consumption
80
no storage
(COP=1)
60
no storage
(COP=3)
40
limited storage
(COP=1)
20
limited storage
(COP=3)
Share off heat
Sh
h
Yearly share, %
100
unlimited
storage (COP=1))
0
0
2000
Wind power (MW)
4000
unlimited
storage (COP=3))
Figure 14. Yearly share of wind power in electricity (black line) and heat consumption (purple,
green and red lines) with different wind power amounts and operating strategies. Reproduced
with permission from Publication IV. Copyright 2015 Elsevier.
7.3.2
Effect of network structure
The importance of network structure and the location of the connection point
are described with a simplified model of the city of Helsinki (Publication III).
Figure 15 shows the spatial consumption profiles and simplified non-looped energy networks used in the simulations. The spatial distribution of consumption
is based on a radial consumption profile of Eq. (1). Different connection points
for a large-scale wind power plant were studied (A–C in Figure 15).
Table 3 summarizes the different cases and their results. The first option considered for the connection point was on the shore side at the end of a major
trunk line of the power grid (point A). At this point, the maximum power channel was around 100 MW, while the heat network could deliver up to 35 MW of
thermal energy by P2H at the same point. If P2H were also used in the downstream locations of point A, the wind power capacity could be increased by 50%
from the non-P2H case to 150 MW.
Next, point B was considered. It is closer to the main junction of the electric
trunk lines, but it is also along the strong line in the heat network. The superior
location in the power grid already triples the wind capacity that is seen without
P2H, but P2H could further increase the wind capacity to more than 1 GW, while
simultaneously producing 11% of the heat demand of the city by P2H. Additionally to point B, point C is a similar node along the main lines of both networks.
Without P2H, these points together would allow for a capacity of more than 600
MW of wind power. With the increased flexibility offered by P2H, the wind
power capacity could be increased to nearly 2 GW.
If heat pumps (COP = 3) were used instead of electric resistance, the wind
capacity could be reduced while maintaining the same share of P2H in heat demand. In the case of point A, the wind capacity would decrease by 8–23 MW
from the values in Table 3.
48
Results for Helsinki
Figure 15. Spatial distribution of electricity (top) and heat (bottom) demands in MW/km 2 for simplified Helsinki simulations. Left: average demand, right: peak demand. Spatial resolution: 1
km × 1 km. Reproduced with permission from Publication III. Copyright 2012 Elsevier.
Table 3. Effects of P2H conversion and the location of the wind power connection on the accessible wind power capacity. A COP of 1 was used in these results. Reproduced from Publication III with permission. Copyright 2012 Elsevier.
Conversion strategy and grid
connection point of wind
power
Wind power at Point A
No P2H conversion
P2H at Point A
P2H downstream of A
Wind power at Point B
No P2H conversion
P2H at Point B
P2H downstream of B
Wind power at Points B & C
No P2H conversion
P2H at Points B & C
Max.
wind
(MW)
Heat by P2H,
peak (MW) /
total annual
(GWh)
Share of
wind in
electricity (%/a)
Share of
wind in
heat
(%/a)
101
135
153
35/0.27
53/9.6
5.3
7.0
7.7
0.037
0.15
319
1070
1080
769/754
779/770
16.7
38.5
38.6
11.4
11.8
609
1943
1367/1293
31.7
71.6
19.6
49
Results for Helsinki
These results demonstrate that the capacity for VRE could be increased when
moving from the city periphery toward the strong grid points and optimal junction points for P2H conversion. P2H offers a major increase in possible VRE
production. Point A could deliver around 8% of the yearly electricity demand in
Helsinki; point B, 39%; and points B and C together, 72%. In all final energy use
(excluding the transport sector), the percentages are 3%, 22% and 40%, respectively.
7.4
Energy system operation with CHP plants and large-scale
wind power (Publication V)
To study the effects of large-scale VRE integration on the existing power plants
of Helsinki, the optimization model of Publication V (Chapter 6 in this thesis)
was used with large wind power schemes. Scenarios of up to 1,500 MW of wind
power were considered, an amount that roughly corresponds to the yearly electricity demand of the city. The option of linking wind power to the heat demand
through a power-to-heat (P2H) conversion was also studied, as doing so increases the flexibility of the system, thus solving the possible mismatch between
power supply and demand.
Three specific cases were analyzed: 1) the reference case with wind power integrated into the power system without any flexibility measures; 2) wind power
with large-scale P2H conversion connected to the district heating system
through electric resistance heating (COP=1); and 3) wind power with P2H connection to the district heating through heat pumps (COP=3).
7.4.1
Parameters for Helsinki energy system
The power plants and their main characteristics are listed in Table 4. Detailed
lists of all the values for technical and economic input parameters are included
in Publication V, but a short summary is given here.
Table 4. Main power plant characteristics used in the Helsinki case. The letters A–F refer to symbols in Figure 8. NG refers to natural gas. Partly reproduced from Publication V with permission. Copyright 2016 Elsevier.
Plant
Electr.
Heat
Min. output
Max ramping
Fuel
(MW)
(MW)
(% of max.)
(%/min)
A Vuosaari CHP
NG
630
580
50
4
B Hanasaari CHP
Coal
220
420
40
3
C Salmisaari CHP
Coal
160
300
40
3
D Heat boilers
NG
–
360
0
4
E Heat boilers
Oil
–
1900
0
5
F Heat boilers
Coal
–
180
0
3
G Katri Vala heat pump
Electr.
–
90
0
–
50
Results for Helsinki
An overall efficiency of 90% was assumed for the CHP plants, and a constant
ratio of power to heat outputs was used. Furthermore, the coefficient of performance of the heat pump plant was also assumed to be constant (COP = 3). In
the P2H cases, the storage capacity of the district heating network is utilized. A
capacity of 5 GWh was assumed, roughly corresponding to a temperature difference of 50 °C in the lines.
The hourly spot price from Nord Pool [85] was used as a selling price for electricity, and the district heating price was taken from the local energy company
Helen [86]. Production costs include fuel costs, operations and maintenance
(O&M) costs, and costs from carbon emissions based on the European emissions trading system (ETS). Fuel costs are from Statistics Finland [87]; for natural gas, a price of 51 €/MWh was used, and for coal, 36 €/MWh. Investment
costs are excluded, since the optimization is done only for the energy system
operation.
7.4.2
Effects of wind power and P2H on plant operation
Figure 16 illustrates the effects of increasing wind power installations on the
traditional CHP production in Helsinki. Rising shares of wind power decrease
the CHP production, as expected. However, in the reference case without P2H,
only gas CHP is being replaced by wind, as the current price ratio of coal and
gas favors the coal-based production over gas CHP. Interestingly, the decrease
in CHP production levels out at high shares of wind. This leveling stems from
the large need of locally produced heat in the city and results in high amounts
of exported electricity (Figure 17). The coal plants have lower power-to-heat ratio and lower fuel price, which makes them preferable in heat production. Furthermore, because of the need for heat, wind power alone would not be able to
decrease the emissions of the city.
When adding flexibility to the system through large-scale heat pump installations, the gas CHP gains an advantage over coal. The P2H “electrifies” the heat
production, which favors the higher power-to-heat ratio of combined-cycle gas
CHP. In addition, the more electric energy system prefers gas CHP due to its
higher flexibility in ramping the output power.
The flexible P2H case with heat pumps thus results in a clearly smaller share
of coal CHP, which also positively affects the CO2 emissions. Using heat pumps
alone (even with no wind power) would decrease the emissions by roughly 30%.
Moreover, decreasing shares of fossil CHP with increasing shares of wind power
would reduce the emissions even further. For example, 1,500 MW of wind together with P2H would drop the emissions by almost 60%. Additionally, the export of electricity would be reduced and the import increased, resulting in a better balance.
51
Power production from exixsting plants
(% of consupmtion)
Results for Helsinki
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0
200
400
600
800
1000 1200 1400 1600
Wind power (MW)
Gas – Reference
Coal – Reference
Gas – COP 1
Coal – COP 1
Gas – COP 3
Coal – COP 3
Figure 16. Electricity produced by CHP plants with wind power and different P2H schemes. Reference: no P2H; COP 1: large-scale P2H with electric resistance heating; COP 3: large-scale
P2H with heat pumps. Reproduced with permission from Publication V. Copyright 2016 Elsevier.
100%
Electricity import/export
(% of consupmtion)
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0
200
400
600
800
1000 1200 1400 1600
Wind power (MW)
Import – Ref
Export – Ref
Import – COP 1
Export – COP 1
Import – COP 3
Export – COP 3
Figure 17. Import and export of electricity with wind power and different P2H schemes. Reference: no P2H; COP 1: large-scale P2H with electric resistance heating; COP 3: large-scale
P2H with heat pumps. Reproduced with permission from Publication V. Copyright 2016 Elsevier.
52
Results for Helsinki
Power production (% of consupmtion)
The use of heat boilers would also be reduced drastically by using P2H. Already, resistance heating (COP=1) and 1,500 MW of wind power would reduce
boiler use by 76–88%. With heat pumps (COP=3), all the boilers could, in theory, be removed. It might, however, be reasonable to leave some boiler capacity
in the city for extreme peak conditions. Reasons for doing so might include a
requirement of higher water temperature in the heat network than can be
achieved with heat pumps, or of a higher capacity factor for P2H (in the modeled
cases between 18–33%). Increasing the heat pump capacity above the modeled
500 MW (electric power) would not further decrease the CHP or boiler operation; rather, it would just lead to increased electricity imports from the spot
market during the low-price hours.
The Emissions Trading System (ETS) includes all large power plants in the
European Union. Thus, the effect of the price of CO2 on the results was analyzed
in this work. Currently, the price is around 10 €/tCO2, and it was compared
against two other prices (0 €/tCO2 and 40 €/tCO2). Figure 18 shows how increasing the emission price shifts production from coal to gas, and this shift occurs much more radically with a higher emission price (40 €/tCO2). However,
at large shares of wind power, the gas plant alone is not able to cope with the
increasing variability in the net demand, and thus the shift from coal to gas is
significantly smaller. With heat pumps, the share of gas CHP decreases with increasing emission price. A large-scale heat pump scheme with an emission price
of 40 €/tCO2 would eliminate almost all coal use in Helsinki.
100%
80%
60%
40%
20%
0%
0
500
1000
Wind power (MW)
Gas – ETS 0
Gas – ETS 10
Gas – ETS 40
1500 0
500
1000
1500
Wind power (MW)
Coal – ETS 0
Coal – ETS 10
Coal – ETS 40
Figure 18. Electricity produced by CHP plants with wind power and emission prices (ETS,
€/tCO2). Left: reference case without P2H; right: P2H with heat pumps (COP=3). Reproduced
with permission from Publication V. Copyright 2016 Elsevier.
53
8. Results for other cities
8.1
Shanghai, China
Shanghai is a megacity in a developing country with rapidly increasing energy
consumption [88]; thus, it is highly important to study this city’s energy system
structure and possibilities for renewable integration in order to avoid further
investments in fossil energy and possibly to reduce its use.
8.1.1
Spatiotemporal profiles with limited input data (Publication II)
The model presented in Chapter 4 and Publication II was also applied to the
case of Shanghai, which offered limited spatial input data. In 2010, the electricity consumption in the city was 130 TWh, of which 55% was the share of industrial loads and 13% was the share of residential loads [89]. The electric load profile of the whole city is shown in Figure 19. It reaches its maximum values during
summers, when the cooling load is high.
Radial density distributions were used in this case. The city is divided into 16
districts, of which the densities of population, industrial employees, and office
floor area are known [90]. Eq. (1) is fitted to these values, and the resulting density profiles are used as an input for the model to produce the spatiotemporal
load profiles.
Figure 19. Electricity load profile of Shanghai, China. Reproduced with permission from Publication III. Copyright Elsevier 2012.
55
Results for other cities
Figure 20. Simulated spatial consumption in Shanghai. Left: average load; right: peak load. Spatial resolution in the figure is 1 km x 1 km. Reproduced with permission from Publication III.
Copyright 2012 Elsevier.
Figure 21. Simulated temporal electricity load profiles at different distances from the Shanghai
city center. Reproduced with permission from Publication III. Copyright 2012 Elsevier.
The resulting spatiotemporal load profiles are shown in Figure 20 and Figure
21. The temporal profiles resemble each other at different distances, because the
consumer classes were assumed to spread throughout the city, and no clear division was made between industrial, residential and office areas. The high share
of industrial load is another reason, as it was assumed to spread even more
evenly than offices (which are concentrated in the city center). The dominance
of industry also explains the lack of clear evening peaks, which are characteristic
of residential areas. The peak power in the city center (slightly over 60
MW/km2) remains lower than in the highest district in Helsinki (80 MW/km2,
Chapter 7.2.2). This is, however, due to the coarse resolution of input data. The
densest district in the city center has an area of 23 km2. If the source data were,
for example, from 1 km2 areas, the peak power would probably exceed 500
MW/km2 in the most high-rise business areas.
56
Results for other cities
8.1.2
Large-scale photovoltaics together with micro-trigeneration (Publication III)
For Shanghai, large-scale distributed photovoltaics (PV) and micro-trigeneration schemes were considered. The detailed analysis and input data are given
in Publication III. The calculations were made with the multi-carrier model of
Publication III (Chapter 5 in this thesis). In Shanghai, the only network in the
model was a power grid, but spatiotemporal consumption profiles were created
for heating and cooling as well. As a starting point, electricity covers 44% of the
heating and 82% of the cooling load.
Results from the simulations are presented in Table 5. In the reference case,
with no PV, electricity must be imported from the large central power plants
close to the city. Case A (in the table) shows the self-use limit of PV-produced
electricity in Shanghai to be 19 GW, which could produce 21% of the electricity
demand in the city. If the PV installations were increased to 150% of the self-use
limit (case B), 31% of the demand could be covered, and, simultaneously, surplus electricity would be produced (in an amount equivalent to 1% of the yearly
electricity load). The surplus could be used in local electric heating systems.
The spatial analysis, however, suggests that the power carrying capacity of the
power grid could be easily exceeded in the periphery areas of the city in this case.
The problems could be solved by rearranging a small portion of the distributed
PV panels closer to the city center with stronger power lines.
Cases C and D include gas-driven trigeneration schemes, which were installed
in the city center area (5 km × 5 km). Trigeneration technology combines gasdriven solid-oxide fuel-cell cogeneration (gas to electricity and heat) and absorption cooling technology (heat to cooling). Assumed efficiencies were 40%
for the gas-to-electricity and gas-to-heat processes and 70% for the heat-tocooling process [91,92]. The trigeneration is operated in such a way that heating
and cooling loads are satisfied first, and then electricity production is adjusted
accordingly. PV installations need to be reduced in order to avoid excess electricity production.
Table 5. Shanghai cases with different PV and trigeneration schemes. DEGS refers to distributed
energy generation systems (i.e., PV + trigeneration). Reproduced from Publication III with
permission. Copyright 2012 Elsevier.
Scheme
A. PV (self-use limit)
B. PV (150% self-use
limit)
C. PV + Trigeneration
covering 25 % of
cooling and heating in city center
D. PV + Tri-generation 50%
PV
GWp
19.3
28.6
PV share
in electricity+, %
21
31
Trigeneration
GW(el)
–
–
DEGS
share in
electricity,
%
21
31
DEGS
share in
energy, %
18
27
16.0
18
3.7
28
28
9.3
11
7.3
31
32
57
Results for other cities
In case D, where 50% of the thermal loads in the city center are satisfied by
trigeneration, it is possible to produce one third of all energy consumption by
distributed means. In addition, the CO2 emissions of the city could be reduced
by 38% assuming a reference system based on coal power. This reduction is due
to the high fuel efficiency of trigeneration technologies.
8.2
Concepción, Chile (Publication VI)
Chile is an interesting country in which to study the integration of solar photovoltaics. There, PV has reached grid parity, and thus the solar PV market is
growing rapidly, especially in urban areas [93]. The city of Concepción was chosen as the case city to study the potential of solar energy as well as the possible
limitations that the power grid infrastructure may impose on the large-scale PV
installations. Such may arise in many developing countries with a high solar resource.
Detailed spatial data about the solar potential in Concepción were presented
in Publication VI. Solar radiation and the urban structure were taken into account in the assessment of the solar potential. With respect to urban structure,
spatial details of the city, such as building typologies and shading due to topography and neighboring buildings, were included in the assessment models.
8.2.1
Electricity consumption, city structure, and the power grid
Energy consumption patterns of the buildings were gathered simultaneously
with the spatial research on building typologies. As the total electricity consumption of the city was also known, it was possible to produce a spatiotemporal
load profile for the city. For this purpose, all the buildings in the city were
grouped into 5500 building combinations. Each of these was then assigned a
load profile depending on the building type (household, service, hospital) and
typology (low, mid, high-rise sparse, and high-rise dense). The electricity consumption profile is shown in Figure 22, along with the PV production curve.
Figure 23 reveals the spatial distribution of the consumption.
Figure 22. Electricity load profile of residential and service sector in Concepción and the PV production curve at the self-use limit of 73 MW. Reproduced with permission from Publication
VI. Copyright 2016 Elsevier.
58
Results for other cities
Accurate data about the power grid were not available, and thus some assumptions were made. The focus was set on the mid-voltage distribution and transmission lines, omitting the low-voltage distribution to customers. The city was
divided into 1 km × 1 km squares, which were connected by the mid-voltage
lines. Load profiles of single building combinations were summed up to form a
single load profile for each square. The squares and assumed power lines are
shown in Figure 23.
Present power demand (without PV) is used to define the power-carrying capacity of the grid. This means that the maximum allowed power flow in a power
line is the flow that would occur in the current situation. This conservative assumption guarantees that no extra stress, compared to the current situation, is
placed on the grid.
Figure 23. Left: Peak power demand in Concepción. Right: Annual power demand with assumed
power grid. Light green shows high-voltage connections between the power stations (light
green dots), and dark green shows the mid-voltage lines. Building combinations are marked
with small blue crosses. Reproduced with permission from Publication VI. Copyright 2016
Elsevier.
8.2.2
Results
The total yearly solar radiation on the buildings of Concepción was found to be
around 3,900 GWh, which exceeds the total energy demand of the city by 22%.
The total solar capacity, however, is severely limited by technical factors (e.g.,
the efficiency of appliances and the capacity of the power grid).
The self-use limit of the PV production in Concepción was found to be 73 MW,
which would cover 13% (95 GWh) of the power consumption. The corresponding production curve is shown in Figure 22. By doubling the amount of PV, 24%
(170 GWh) of the consumption could be satisfied, but 17 GWh of surplus electricity would be produced.
PV production throughout the city decreases the consumer’s need to buy
power from the grid. Therefore, with relatively low amounts of PV (1–2 × selfuse limit) the average power flows and, thus, the stress in the grid decrease (Figure 24). No overflow problems to the grid are caused by 150 MW of PV, but the
momentary oversupply of PV production turns the flow from downstream to
upstream. The upstream flow, however, stays below the maximum downstream
flow of the reference case.
In the case of increasing the PV installations closer to the total potential of the
city, overflow problems will occur. For example, PV of five times the self-use
59
Results for other cities
limit (i.e., 370 MW) corresponds to over 80% of the total solar capacity (assuming 15% efficiency for PV), and would produce 34% of the electricity demand,
but the capacity of the grid would be exceeded by up to 80% in some lines. This
is shown in Figure 25. In practice, all parts of the grid would be overstressed.
The problem could be solved, for example, by local P2H appliances. If all excess
solar PV production were converted to heat, about 28% of the heat demand
could be satisfied.
The high-PV case, also demonstrates how the role of the grid changes from
simply supplying power into a way of balancing fluctuating PV production. The
average flows in Figure 25 (right) are in the same range as in the reference case
with no PV (Figure 24, left), but the maximum peaks are clearly stronger (Figure
25, left).
Figure 24. Average power flows in the grid lines in Concepción as a percentage of the power line
size. Left: reference case with no PV; right: 150 MW of PV (2 × self-use limit). Reproduced
with permission from Publication VI. Copyright 2016 Elsevier.
Figure 25. Power flows in the grid lines in Concepción with 370 MW of PV (5 × self-use limit).
Left: maximum flows; right: average flows as a percentage of the power line size. The color
bars have different scales. Reproduced with permission from Publication VI. Copyright 2016
Elsevier.
60
9. Conclusions
The use of variable renewable electricity, such as wind and solar power, is increasing rapidly around the world. The use of such renewable energy sources
will increase the unpredictability and fluctuations in the net load patterns. Energy systems need to match the energy supply to the demand. Traditionally,
controllable combustion power plants have handled the problem, but with
growing amounts of VRE, new flexibility measures are required. On the other
hand, the role of urban areas in climate change mitigation will increase due to
the global urbanization. The aim of this work was to study the characteristics of
urban energy systems and their readiness to incorporate large-scale VRE
schemes. For the purpose, three novel energy system models were created.
The first model focused on the spatiotemporal patterns of energy consumption
in cities. This type of data is seldom available with sufficient spatial resolution,
and previous models rarely consider both the space and time dimensions. The
developed model, which is based on readily available geographical data, is able
to reach satisfactory spatial and temporal resolutions, thus enabling detailed
analyses of the urban energy systems.
The second model analyzes the possible advantages of viewing the energy system as a whole, taking all final energy carriers into account. Traditionally, energy carriers are handled one at a time, thus omitting several degrees of freedom
in energy system flexibility. In the model, a city is divided into nodes of distributed energy production, consumption, and mutual conversion between energy
carriers. The nodes are connected with different energy carrier networks. This
work focused on electric power grid and district heating, but the model allows
generic multi-carrier analyses. Another advantage over the existing literature is
the comprehensiveness of the model, which simultaneously considers both
macro and micro levels, and includes control functions and network intelligence.
The third question addressed by this work involved the role of the existing
power plants in the new situation with increased VRE generation. A new optimization model is thus presented, which simulates the most economical use of
energy system assets. It is based on a mixed-integer linear problem, and it can
handle many types of energy systems, although this work focuses on the city
scale. Different flexibility measures are included in the model, such as powerto-heat conversion, thermal energy storage, and electricity spot markets. Compared to most of the existing models, the model is kept simple and is fast to run
and easy to modify, despite the fact that it covers the whole energy system, from
61
Conclusions
different energy carriers to several flexibility technologies and spot markets. Another advantage of VRE dynamics is the shorter time step of the model (e.g., 10
minutes), instead of the traditional hourly or even longer time step.
In the case of Helsinki, Finland, the analysis focused on how large-scale wind
power schemes could be integrated into a mid-sized northern city with a high
heating load and a well-established district heating network that covers the
whole city area. The self-use limit of wind power was found to be around 400
MW, which could satisfy up to 27% of the electricity consumption of the city.
With power-to-heat conversion and thermal storage, however, this number
could be dramatically increased. For example, 2000 MW of wind power would
be able to cover more than 60% of the electricity consumption and 20% of the
heat consumption of the city if P2H with electric resistance were used. With heat
pumps, the share in heat could theoretically be increased to close to 70%, but
this increase could be limited by the heating network capacity, and thus requiring large-scale heat storage.
The optimization results clearly indicate the need for system flexibility in the
wind power integration in Helsinki. Without any flexibility measures, wind
power alone would not be able to replace the fossil-based CHP production. Most
of the wind power production would thus be exported outside the city, mainly
due to the high demand for heating and the need to produce it locally. P2H with
heat pumps would change the situation by electrifying the heat production, thus
resulting in decreasing coal use when wind power is increased. Sensitivity analysis against fuel and emission (European emissions trading system) prices show
that P2H makes the coal-based CHP more sensitive with respect to price increases, thus offering considerable help in eliminating the use of coal in the energy production of Helsinki.
The co-operation of different energy carriers was also studied in the southern
megacity of Shanghai, China. The self-use limit for solar PV integration was
found to be below 20 GW (accounting for 20% of the power demand). In Shanghai, a distributed trigeneration scheme was studied with fuel cells and absorption chillers to produce electricity, heat, and cooling. Trigeneration, together
with PV, was able to cover more than 30% of all electricity and energy consumption in the city, yielding a savings of almost 30 Mt of CO2.
The city of Concepción, Chile, is a good example of the importance of studying
limitations rising from the energy system. Even though the overall solar potential, in theory, exceeds the energy demand of the buildings, technical efficiency
and power grid infrastructure limits the PV share in electricity to about 24%.
However, with a local P2H strategy, this could be increased to at least 34%, and,
simultaneously, 28% of heat would be produced without overloading the electric
power grid. Thus, to form a realistic estimate of the solar energy possibilities,
both the pure VRE potential and the power system infrastructure need to be
assessed.
The presented results show the importance of system-wide thinking when defining policies for energy system transition and carbon neutrality. Energy systems must be considered as a whole when flexibility is sought. It was observed
that inherent features in energy systems, such as smart use of district heating
62
Conclusions
networks, could offer much of the needed flexibility for large-scale renewable
energy integration. Therefore, massive investments in traditional network infrastructure could largely be avoided.
For example, viewing electric and thermal systems together instead of as two
separate systems could reveal remarkable new options for VRE integration. P2H
technologies are able to at least double the VRE intake of energy systems, but
with district heating networks and thermal storage the increase could be up to
fivefold. These results show examples of how smart design and systematic
thinking can turn renewable technologies into a mainstream option for energy
production in cities.
The developed models did not include investment analysis tools, as the focus
was on the operation of energy systems. Therefore, in the future, it would be
interesting to integrate investment options into the models. Another interesting
update would be to expand the analyses to the national or even continental
scale. This would, however, require fundamental market models to be integrated.
63
References
[1]
Olivier JGJ, Janssens-Maenhout G, Muntean M, Peters JAHW. Trends in
Global CO2 Emissions: Report 2015. The Hague: 2015.
[2]
IPCC. Climate Change 2014: Mitigation of Climate Change. Contribution of
Working Group III to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge, UK and New York, USA: Cambridge
University Press; 2014.
[3]
WEC. World Energy Scenarios: Composing energy futures to 2050. London,
UK: 2013.
[4]
IRENA. Roadmap for a Renewable Energy Future. 2016 ed. Abu Dhabi, UAE:
International Renewable Energy Agency (IRENA); 2016.
[5]
IEA International Energy Agency. Solar Energy Perspectives. Paris, France:
2011.
[6]
Freris L, Infield D. Renewable Energy in Power Systems. Wiley; 2008.
[7]
Gostling J. Two Shifting of Power Plant: Damage to Power Plant Due to Cycling
– A brief overview. Ommi 2002;1:1–9.
[8]
Ueckerdt F, Hirth L, Luderer G, Edenhofer O. System LCOE: What are the
costs of variable renewables? Energy 2013;63:61–75.
doi:10.1016/j.energy.2013.10.072.
[9]
ARUP. Five minute guide to Energy in Cities. 2016.
[10]
United Nations Department of Economic and Social Affairs: Population
Division. World Urbanization Prospects: The 2011 Revision, CD-ROM Edition.
2012.
[11]
Pepermans G, Driesen J, Haeseldonckx D, Belmans R, D’haeseleer W.
Distributed generation: definition, benefits and issues. Energy Policy
2005;33:787–98. doi:10.1016/j.enpol.2003.10.004.
[12]
WBCSD (World Business Council for Sustainable Development). Energy
Efficiency in Buildings – Transforming the Market. Switzerland: 2009.
[13]
IEA (International Energy Agency). Worldwide Trends in Energy Use and
Efficiency. Paris, France: 2008.
[14]
Ma Q, Luo L, Wang RZ, Sauce G. A review on transportation of heat energy
over long distance: Exploratory development. Renewable and Sustainable
Energy Reviews 2009;13:1532–40. doi:10.1016/j.rser.2008.10.004.
[15]
The World Bank. World Development Indicators 2016.
[16]
Klimstra J. Power Supply Challenges. 2nd ed. Vaasa: Wärtsilä Finland Oyj;
2014.
[17]
U.S. Energy Information Administration. Energy in Brief: What is the electric
power grid and what are some challenges it faces? 2015.
http://www.eia.gov/energy_in_brief/article/power_grid.cfm (accessed August
24, 2016).
[18]
Fingrid. Power System in Finland 2016.
http://www.fingrid.fi/en/powersystem/general description/Power System in
Finland/Pages/default.aspx (accessed August 24, 2016).
[19]
Short TA. Electric Power Distribution Handbook. Boca Raton, Florida, USA:
65
References
CRC Press; 2014.
[20]
Niemi R, Lund PD. Alternative ways for voltage control in smart grids with
distributed electricity generation. International Journal of Energy Research
2012:1032–43. doi:10.1002/er.1865.
[21]
Ipakchi A, Albuyeh F. Grid of the future. IEEE Power and Energy Magazine
2009;7:52–62. doi:10.1109/MPE.2008.931384.
[22]
Farhangi H. The path of the smart grid. IEEE Power and Energy Magazine
2010;8:18–28. doi:10.1109/MPE.2009.934876.
[23]
American Council for an Energy-Efficient Economy. Types of Heating Systems.
SmarterHouse.org 2015. http://smarterhouse.org/heating-systems/typesheating-systems (accessed August 24, 2016).
[24]
Laughton C. Solar Domestic Water Heating. New York, USA: Earthscan; 2010.
[25]
Lechner N. Heating, Cooling, Lighting: Sustainable Design Methods fro
Architects. 4th ed. Hoboken, New Jersey: Wiley; 2015.
[26]
Finnish Energy. District Heat on a Global Scale 2016.
http://energia.fi/en/energy-and-environment/district-heat-and-districtcooling/district-heat-global-scale (accessed August 24, 2016).
[27]
Vattenfall. About District Heating 2014.
https://corporate.vattenfall.com/about-energy/energy-distribution/districtheating/ (accessed August 24, 2016).
[28]
Koskelainen L, Saarela R, Sipilä K. Kaukolämmön käsikirja. Helsinki, Finland:
Energiateollisuus ry; 2006.
[29]
Hepbasli A. Thermodynamic analysis of a ground-source heat pump system for
district heating. International Journal of Energy Research 2005;29:671–87.
doi:10.1002/er.1099.
[30]
Schmidt T, Mangold D, Müller-Steinhagen H. Central solar heating plants with
seasonal storage in Germany. Solar Energy 2004;76:165–74.
doi:10.1016/j.solener.2003.07.025.
[31]
Lindenberger D, Bruckner T, Groscurth H-M, Kümmel R. Optimization of solar
district heating systems: seasonal storage, heat pumps, and cogeneration.
Energy 2000;25:591–608. doi:10.1016/S0360-5442(99)00082-1.
[32]
Reddy TB, Linden D. Linden’s Handbook of Batteries. 4th ed. New York:
McGraw-Hill; 2011.
[33]
U.S. Department of Energy. Ventilation Systems for Cooling 2016.
http://energy.gov/energysaver/ventilation-systems-cooling.
[34]
Wang SK. Handbook of Air Conditioning and Refrigeration. New York, USA:
McGraw-Hill; 2001.
[35]
Srikhirin P, Aphornratana S, Chungpaibulpatana S. A review of absorption
refrigeration technologies. Renewable and Sustainable Energy Reviews
2001;5:343–72. doi:10.1016/S1364-0321(01)00003-X.
[36]
Kim DS. Solar absorption cooling. TU Delft, 2007.
[37]
Söderman J. Optimisation of structure and operation of district cooling
networks in urban regions. Applied Thermal Engineering 2007;27:2665–76.
doi:10.1016/j.applthermaleng.2007.05.004.
[38]
Chow TT, Fong KF, Chan ALS, Yau R, Au WH, Cheng V. Energy modelling of
district cooling system for new urban development. Energy and Buildings
2004;36:1153–62. doi:10.1016/j.enbuild.2004.04.002.
[39]
Klimstra J, Hotakainen M. Smart Power Generation. 3rd ed. Helsinki, Finland:
Avain Publishers; 2011.
[40]
Cochran J, Miller M, Zinaman O, Milligan M, Arent D, Palmintier B, et al.
66
References
Flexibility in 21st Century Power Systems 2014.
[41]
Corsi S. Voltage Control and Protection in Electrical Power Systems: From
System Components to Wide-area Control. London, UK: Springer-Verlag;
2015.
[42]
Lund PD, Lindgren J, Mikkola J, Salpakari J. Review of energy system
flexibility measures to enable high levels of variable renewable electricity.
Renewable and Sustainable Energy Reviews 2015;45:785–807.
doi:10.1016/j.rser.2015.01.057.
[43]
Kirby B. Ancillary Services: Technical and Commercial Insights. 2007.
[44]
North American Electric Reliability Corporation. Ancillary Service and
Balancing Authority Area Solutions to Integrate Variable Generation.
Princeton, New Jersey: 2011.
[45]
Aalto A, Honkasalo N, Järvinen P, Jääskeläinen J, Raiko M, Sarvaranta A.
Mistä lisäjoustoa sähköjärjestelmään? Loppuraportti. Espoo: 2012.
[46]
Burke DJ, O’Malley MJ. Factors Influencing Wind Energy Curtailment. IEEE
Transactions on Sustainable Energy 2011;2:185–93.
doi:10.1109/TSTE.2011.2104981.
[47]
IEA International Energy Agency. Harnessing Variable Renewables: a Guide to
the Balancing Challenge. Paris, France: IEA; 2011.
[48]
Riesz J, Milligan M. Designing electricity markets for a high penetration of
variable renewables. Wiley Interdisciplinary Reviews: Energy and Environment
2015;4:279–89. doi:10.1002/wene.137.
[49]
Denholm P, Ela E, Kirby B, Milligan M. The Role of Energy Storage with
Renewable Electricity Generation. Golden, Colorado: 2010.
[50]
Hadjipaschalis I, Poullikkas A, Efthimiou V. Overview of current and future
energy storage technologies for electric power applications. Renewable and
Sustainable Energy Reviews 2009;13:1513–22. doi:10.1016/j.rser.2008.09.028.
[51]
Kondoh J, Ishii I, Yamaguchi H, Murata A, Otani K, Sakuta K, et al. Electrical
energy storage systems for energy networks. Energy Conversion and
Management 2000;41:1863–74. doi:10.1016/S0196-8904(00)00028-5.
[52]
Tuohy A, Kaun B, Entriken R. Storage and demand-side options for integrating
wind power. Wiley Interdisciplinary Reviews: Energy and Environment
2014;3:93–109. doi:10.1002/wene.92.
[53]
Hall PJ, Bain EJ. Energy-storage technologies and electricity generation.
Energy Policy 2008;36:4352–5. doi:10.1016/j.enpol.2008.09.037.
[54]
Ferreira HL, Garde R, Fulli G, Kling W, Lopes JP. Characterisation of electrical
energy storage technologies. Energy 2013;53:288–98.
doi:10.1016/j.energy.2013.02.037.
[55]
Ummels BC, Pelgrum E, Kling WL. Integration of large-scale wind power and
use of energy storage in the Netherlands’ electricity supply. IET Renewable
Power Generation 2008;2:34. doi:10.1049/iet-rpg:20070056.
[56]
Waite M, Modi V. Potential for increased wind-generated electricity utilization
using heat pumps in urban areas. Applied Energy 2014.
doi:10.1016/j.apenergy.2014.04.059.
[57]
Specht M, Baumgart F, Feigl B, Frick V, Stürmer B, Zuberbühler U, et al.
Storing bioenergy and renewable electricity in the natural gas grid. FVEE AEE
Topics 2009 2009:69–78.
[58]
Gahleitner G. Hydrogen from renewable electricity: An international review of
power-to-gas pilot plants for stationary applications. International Journal of
Hydrogen Energy 2013;38:2039–61. doi:10.1016/j.ijhydene.2012.12.010.
[59]
North American Electric Reliability Corporation. Potential Reliability Impacts
67
References
of Emerging Flexible Resources. Princeton, New Jersey: 2010.
[60]
Barter P. Reinventing Parking: “Cars are parked 95% of the time”. Let’s check!
2013. http://www.reinventingparking.org/2013/02/cars-are-parked-95-oftime-lets-check.html (accessed May 19, 2016).
[61]
Gellings C, Smith W. Integrating demand-side management into utility
planning. Proceedings of the IEEE 1989;77:908–18.
[62]
Cappers P, Mills A, Goldman C, Wiser R, Eto JH. Mass Market Demand
Response and Variable Generation Integration Issues: A Scoping Study.
Berkeley, California: 2011.
[63]
Kirschen D. Demand-side view of electricity markets. Power Systems, IEEE
Transactions on 2003;18:520–7.
[64]
Mathieu J, Haring T, Andersson J. Residential demand response program
design: Engineering and Economic perspectives. European Energy Markets
(EEM) Conference, Stockholm, Sweden: 2013.
[65]
Nourai A, Kogan VI, Schafer CM. Load Leveling Reduces T&D Line Losses.
IEEE Transactions on Power Delivery 2008;23:2168–73.
doi:10.1109/TPWRD.2008.921128.
[66]
O’Dwyer C, Duignan R, O’Malley M. Modeling demand response in the
residential sector for the provision of reserves. Power and Energy Society
General Meeting, 2012 IEEE, San Diego, CA, USA: 2012.
doi:10.1109/PESGM.2012.6344757.
[67]
Schaber K, Steinke F, Hamacher T. Transmission grid extensions for the
integration of variable renewable energies in Europe: Who benefits where?
Energy Policy 2012;43:123–35. doi:10.1016/j.enpol.2011.12.040.
[68]
Macilwain C. Supergrid. Nature 2010;468:624–5.
[69]
Bahrman MP, Johnson BK. The ABCs of HVDC transmission technologies.
IEEE Power and Energy Magazine 2007;5:32–44.
doi:10.1109/MPAE.2007.329194.
[70]
Lopes JAP, Madureira AG, Moreira CCLM. A view of microgrids. Wiley
Interdisciplinary Reviews: Energy and Environment 2013;2:86–103.
doi:10.1002/wene.34.
[71]
Lasseter RH, Paigi P. Microgrid: a conceptual solution. 2004 IEEE 35th Annual
Power Electronics Specialists Conference (IEEE Cat. No.04CH37551), IEEE;
2004, p. 4285–90. doi:10.1109/PESC.2004.1354758.
[72]
Tanner JC. Factors affecting the amount of travel. Road Research Technical
Papers 1961;51:46–73.
[73]
Kulworawanichpong T. Simplified Newton–Raphson power-flow solution
method. International Journal of Electrical Power & Energy Systems
2010;32:551–8. doi:10.1016/j.ijepes.2009.11.011.
[74]
Gardiner CW. Handbook of Stochastic Methods for Physics, Chemistry and the
Natural Sciences. 2nd ed. Berlin: Springer-Verlag; 1985.
[75]
Helen Ltd. Helen investoi uusiutuvaan energiaan ja hajautettuun
energiantuotantoon 2015. https://www.helen.fi/uutiset/2015/helen-investoiuusiutuvaan-energiaan-ja-hajautettuun-energiantuotantoon/ (accessed June 8,
2016).
[76]
Helsinki Environment Centre. Helsinki Environmental Statistics 2015.
helsinginymparistotilasto.fi (accessed August 26, 2016).
[77]
Helsingin Energia. Helsingin Energialle kansainvälinen ympäristöpalkinto
2013. https://www.helen.fi/uutiset/2013/helsingin-energialle-kansainvalinenymparistopalkinto/ (accessed June 10, 2015).
[78]
Finnish Energy Industries. District heating in Finland 2013. 2014.
68
References
[79]
Helen Ltd. Origin of energy. About Us – Energy Production 2016.
https://www.helen.fi/en/helen-oy/about-us/energy-production/origin-ofenergy/ (accessed June 9, 2016).
[80]
City of Helsinki Urban Facts. Helsinki alueittain 2012 (Helsinki by District
2012). Helsinki, Finland: 2013.
[81]
E.ON Mitte AG. Standardlastprofile 2011. Accessed: Sept 16, 2013:
http://www.eonmitte.com/de/netz/veroeffentlichungen/strom_/standardlastprofil_verfahren
/standardlastprofile_2011.
[82]
Helsinki Environment Centre. Helsinki Environmental Statistics 2013.
http://www.helsinginymparistotilasto.fi/ (accessed August 25, 2015).
[83]
City of Helsinki Urban Facts, Statistics Finland. Helsingin seudun aluesarjat (in
Finnish) 2013:http://www.aluesarjat.fi. www.aluesarjat.fi.
[84]
Lund PD. Clean energy systems as mainstream energy options. International
Journal of Energy Research 2015. doi:10.1002/er.3283.
[85]
Nordpool Spot. Historical Market Data 2013.
http://www.nordpoolspot.com/historical-market-data/ (accessed July 24,
2015).
[86]
Helen Ltd. Kaukolämmön energia- ja vesivirtamaksut. 2015.
[87]
Statistics Finland. Fuel classification 2015 2015.
http://tilastokeskus.fi/tup/khkinv/khkaasut_polttoaineluokitus.html
(accessed October 10, 2015).
[88]
Ruet J, Vallantin F, Daval A, Pasternak J. WEC « Energy for Megacities »
Study – Shanghai Municipality Case Study.
[89]
Department of Energy Statistics – National Bureau of Statistics of the People’s
Republic of China. China Energy Statistical Yearbook 2011. China Statistics
Press; 2011.
[90]
Shanghai Municipal Statistics Bureau. Shanghai Statistical Yearbook 2012.
China Statistics Press; 2012.
[91]
Hawkes A, Staffell I, Brett D, Brandon N. Fuel cells for micro-combined heat
and power generation. Energy & Environmental Science 2009;2:729–44.
doi:10.1039/b902222h.
[92]
Absorption Chillers 2012.
http://www.cogeneration.net/Absorption_Chillers.htm (accessed February 6,
2012).
[93]
Campos P, Troncoso L, Lund PD, Cuevas C, Fissore A, Garcia R. Potential of
distributed photovoltaics in urban Chile. Solar Energy 2016;135:43–9.
doi:10.1016/j.solener.2016.05.043.
69
Aalto-DD 16/2017
9HSTFMG*ahcggb+
ISBN 978-952-60-7266-1 (printed)
ISBN 978-952-60-7265-4 (pdf)
ISSN-L 1799-4934
ISSN 1799-4934 (printed)
ISSN 1799-4942 (pdf)
Aalto Universit y
Sc hool of Scienc e
Depar tment of Applied Physics
www.aalto.fi
BUSINESS +
ECON OMY
AR T +
DESIGN +
ARC HITECTURE
SCIEN CE +
TECHNOLOGY
CROSSOVER
D O C TO R A L
DISSERTATIONS