MARKET AND ENVIRONMENT DRIVEN OPTIMIZATION OF FLEXIBLE DISTRIBUTED MULTI-GENERATION

MARKET AND ENVIRONMENT DRIVEN OPTIMIZATION OF FLEXIBLE DISTRIBUTED MULTI-GENERATION
FACULTY OF ELECTRICAL ENGINEERING AND COMPUTING
TOMISLAV CAPUDER
MARKET AND ENVIRONMENT DRIVEN OPTIMIZATION
OF FLEXIBLE DISTRIBUTED MULTI-GENERATION
DOCTORAL THESIS
Zagreb, 2014
FACULTY OF ELECTRICAL ENGINEERING AND COMPUTING
TOMISLAV CAPUDER
MARKET AND ENVIRONMENT DRIVEN OPTIMIZATION
OF FLEXIBLE DISTRIBUTED MULTI-GENERATION
DOCTORAL THESIS
Advisor: Professor Davor Škrlec, PhD
Zagreb, 2014
FAKULTET ELEKTROTEHNIKE I RAČUNARSTVA
TOMISLAV CAPUDER
TRŽIŠNO I OKOLIŠNO UVJETOVANA OPTIMIZACIJA
FLEKSIBILNIH DISTRIBUIRANIH
VIŠEGENERACIJSKIH JEDINICA
DOKTORSKI RAD
Mentor: Prof. dr. sc. Davor Škrlec
Zagreb, 2014.
Doctoral thesis was defended at the University of Zagreb Faculty of electrical Engineering and
Computing, at the Department of Energy and Power Systems.
Advisor: Professor Davor Škrlec, PhD
Doctoral Thesis has 121 pages.
Doctoral thesis number: _______
Advisor’s Biography
Davor Škrlec was born in Zagreb on 1st January, 1963. He graduate at the University of Zagreb
Electrotechnical Faculty in 1986. in the field of Electrical Power Engineering and in the field of
Nuclear Power Engineering. He received M.Sc. and Ph.D. degrees in electrical engineering
(mentor prof. Slavko Krajcar) from the University of Zagreb, Faculty of Electrical Engineering
and Computing (FER), Zagreb, Croatia, in 1990 and 1996 respectively.
From January 1987 he is working at the Department of Energy and Power Systems at FER. From
Janury 2012 until July 2012 he was assistant minister in the Ministry of Environmental and
Nature Protection. In March 2012 he was promoted to Full Professor with tenure.
He participated in 5 scientific projects financed by the Ministry of Science, Education and Sports
of the Republic of Croatia , he was project coordinator in one industrial research project and one
international project and participated in two bilateral research projects. Currently he is a involved
in 2 EU FP7 projects. He published more than 80 papers in journals and conference proceedings
in the area of planning and operation of electrical networks, integration of renewable and
distributed resources, and application of geoinformation systems in power engineering.
Professor Škrlec is a member of IEEE, CIGRE, CIRED and Croatian Nuclear Society. He
participated in 5 conference international programs committees, he is member of a journal
editorial board and he serves as a technical reviewer for various international journals and
conferences. He received silver medal "Josip Lončar" from FER for outstanding Ph.D. theses in
1996.
Životopis mentora
Davor Škrlec je rođen 1. siječnja 1963. godine u Vinkovcima, Republika Hrvatska. Na
Elektrotehničkom fakultetu u Zagrebu diplomirao je 1986. godine na smjeru Energetika, i na
smjeru Nuklearna energetika. Magistrirao je i doktorirao u polju elektrotehnike (mentor
prof.dr.sc. Slavko Krajcar) na Sveučilištu u Zagrebu Fakultetu elektrotehnike i računarstva,
1990. odnosno 1996. godine.
Od siječnja 1987. godine zaposlen je na Zavodu za visoki napon i energetiku FER-a. Od siječnja
do srpnja 2012. godine bio je pomoćnik ministra u Ministarstvu zaštite okoliša i prirode. U
ožujku 2012. godine izabran je za redovitog profesora u trajnom zvanju.
Sudjelovao je na pet znanstvenih projekata Ministarstva znanosti, obrazovanja i sporta Republike
Hrvatske, na jednom od njih voditelj, i bio je voditelj domaćeg industrijskog znanstvenog
projekta. Bio je voditelj jednog međunarodnog projekta i suradnik na dva međunarodna
bilateralna projekta. Trenutačno je suradnik na dva FP7 projekta (Smartgrids ERA-Net i
ACROSS). Objavio je više od 80 radova u časopisima i zbornicima konferencija iz područja
planiranja i pogona elektroenergetskih mreža, integracije obnovljivih izvora energije i
distribuirane proizvodnje u elektroenergetski sustav, te primjenu geoinformacijskih sustava u
elektrotehnici i računarstvu.
Profesor Škrlec je član je nekoliko stručnih organizacija u kojima obavlja ili je obavljao upravne
funkcije: IEEE Hrvatska sekcija, HRO CIGRE, HO CIRED, Hrvatsko nuklearno društvo.
Sudjeluje u 5 međunarodnih programskih odbora znanstvenih konferencija, član je uredničkog
odbora znanstvenog časopisa te sudjeluje kao recenzent u većem broju inozemnih časopisa i
konferencija. Odlukom Fakultetskog vijeća od 18. prosinca 1996. godine dodijeljena mu je
srebrna plaketa "Josip Lončar" za posebno istaknutu doktorsku disertaciju.
Abstract
The share of renewable energy sources in electric power systems around the world is rapidly
increasing and this trend is expected to continue in the future. According to "EU Energy and
Climate Package" the European Union has set ground for future strategies indicating a goal of
20% share of renewable energy sources in total energy mix by the year 2020, together with 20%
more efficient energy use, 20% reduced CO2 emissions and 10% more efficient transport. Several
reports and documents indicate that development of Smart Grids and related technologies will be
a key component in successful integrations of renewables. However, majority of the current
energy system strategies, even in the Smart Grid concept, decouple different energy vectors and
propose operational and strategic development schemes independent of each other. This in
particular relates to cities as the number of people living in the cities is constantly increasing
(today almost 70% of Europe’s population is living in the cities) and are producing around 80%
of global carbon emissions. To this, energy is contributing with 75%. In the above mentioned
context, modern cities need to be looked at as energy hubs where strategic planning decisions are
a multilayer problem describing interactions between different infrastructures and multiple
energy vectors. Capturing interactions between different infrastructures and taking into account a
number of relevant factors needs to be recognized as a necessity for environmental and economic
analyses in sustainable multi-energy system development.
Electricity sector decarbonising strategies, supporting integration of renewable energy sources,
are followed by the similar actions in heating, displacing carbon emissions usually by electrifying
heat. In addition, growing environmental awareness and strong political support have promoted
electric vehicles (EV) as an attractive means of transportation. This in turn means that two energy
sectors, majorly contributing to overall CO2 emissions are directing their future decarbonising
strategies towards electrification. The question is will future heat sector, based on electric heat
pumps and electric resistors, and the electrified transport system in form of increasing EV
integration, bring additional value in achieving goals of reducing CO2 emissions, or will the new
passive electric load just further increase this issue. Opponents of integrating renewable energy
sources frequently point out the uncertainty and variability of electricity production from
renewable energy sources resources and difficulties related to planning and accurately predicting
their output in time. A large share of variable production will inevitably cause changes in the
electric power system control and dynamics and if concepts to system operation and planning do
not change, electrification actions could have the opposite effect from the desired one.
Distributed Generation is seen by many as the electricity production paradigm of the new
millennium. In recent years, the Distributed Multi-Generation (DMG) concept has emerged as a
concept of operation and planning beyond electricity only approach, whereby DG technologies
based on thermal prime movers are capable of increasing the overall generation efficiency by
producing manifold forms of energy, such as electricity, heat, cooling. Distributed MultiGeneration has the potential to provide primary energy saving and emission reduction relative to
conventional separate generation. The power/energy a single distributed generation unit can
provide to the system is rather small, thus there is a need for an aggregated market subject called
a Virtual Power Plant (VPP), which is linking-up multiple small distributed power sources like
wind turbines, combined heat and power (CHP) units, photovoltaic systems and flexible loads.
Centralized control and participation in electricity market as one entity can bring benefits in
terms of providing different services and better forecasting due to generation and consumption
aggregation.
On the above premise, the aim of this thesis is to define a comprehensive and unified technoeconomic and environmental modelling and optimization framework for the operational and
planning evaluation of different DMG options for district energy systems. The concept of
flexibility, seen here from the twofold point of view of operational flexibility (capability to
respond to price signals in close to real time) and planning flexibility (robustness to evolving
scenarios and capability to adapt to changing market conditions), will be analyzed in detail for
the different options.
Keywords: cogeneration, distributed multi-generation, electric heat pumps, flexibility, multienergy systems, virtual power plant, thermal storage
Sažetak
"Tržišno i okolišno uvjetovana optimizacija fleksibilnih distribuiranih višegeneracijskih jedinica"
Sigurnost opskrbe energijom danas još uvijek velikim dijelom ovisi o dostupnosti osnovnih
energenata, fosilnih goriva. Istovremeno projekcije potrošnje električne energije predviđaju
utrostručenje ukupne potrošnje energije u svijetu do 2050. godine. Upravo iz tih razloga, u
dokumentu "EU Energy and Climate Package", razvojna energetska politika Europske unije
postavila je ciljeve: 20% udjela obnovljivih izvora energije, 20% povećanu učinkovitost
korištenja energije, 20% manje emisije CO2 i 10% veću učinkovitost prometnog sektora do 2020.
godine.
Danas je udio obnovljivih izvora u ukupnoj proizvodnji oko 3%, no pretpostavlja se da će se taj
postotak udvostručavati svake tri godine. Kao jedan od ključnih čimbenika u uspješnoj integraciji
visokog udjela obnovljivih izvora energije promovira se implementacija koncepata naprednih
energetskih mreža - Smart Grida. Većina energetskih strategija danas, čak i u konceptu Smart
Grida, razdvaja planiranja i strategije energetskih sektora, ne prepoznajući važnost i prednosti
njihove interakcije. Posebno se to odnosi na gradove, u kojima danas u Europi živi preko 70%
populacije, koji proizvode oko 80% globalnih emisija stakleničkih plinova. Stoga moderne
gradove treba razmatrati kao energetska čvorišta u kojim se odvijaju interakcije svih energetskih
vektora, te njihove strategije planiranja i razvoja moraju usvojiti multidisciplinarni pristup.
Jedino obuhvaćanjem velikog broja relevantnih faktora, koji adekvatno opisuju uzročno
posljedične povezanosti, moguće je provesti kvalitetnu strategiju održivog razvoja za budućnost.
Strategije usmjerene na dekarbonizaciju elektroenergetskog sektora prate slične smjernice u
toplinskom sektoru, gdje se koncepti smanjenja CO2 emisija uglavnom usmjeravaju prema
elektrifikaciji grijanja. Razvijena ekološka svijest istovremeno utječe na sve veće usmjeravanje
prometnog sektora prema električnim vozilima kao učinkovitom, ekološkom i atraktivnom
načinu prijevoza u budućnosti. Posljedično, to znači da dva važna sektora, koji doprinose
ukupnim CO2 emisijama preko 70%, usmjeravaju svoje strategije prema elektrifikaciji. Postavlja
se pitanje hoće li budući toplinski sektor, temeljen na dizalicama topline i električnim grijačima
te promet usmjeren električnim vozilima, doprinijeti smanjenju stakleničkih plinova ili će se
ponašati kao dodatni pasivni teret i još više produbiti problem. Pitanje je na koji način će se takve
promjene reflektirati na budući elektroenergetski sektor.
Protivnici integracije obnovljivih izvora energije naglašavaju promjenjivu i nepredvidljivu
prirodu proizvodnje kao prepreku dobrom vođenju budućih elektroenergetskih sustava; s
razlogom, jer veći broj obnovljivih izvora energije neminovno traži promjenu koncepata vođenja
sustava, kako učinak njihove integracije ne bi bio upravo suprotan željenom. Pravo pitanje je koje
su to promjene potrebne i na kojim razinama se moraju provesti kako bi se mogli uspješno
ostvarili zacrtani ciljevi.
Distribuiranu proizvodnju mnogi vide kao paradigmu proizvodnje električne energije u novom
tisućljeću. U posljednjih nekoliko godina koncept distribuiranih jedinica nadišao je razinu
proizvodnje električne energije, pri čemu su tehnologije distribuirane proizvodnje koje se temelje
na unutrašnjem izgaranju sposobne povećati ukupnu učinkovitost proizvodnjom više oblika
korisne energije poput električne energije te energije grijanja i hlađenja. Pritom se iskorištava
otpadna toplina (na primjer u klasičnim kogeneracijskim postrojenjima) i/ili međusobnim
povezivanjem takvih jedinica kao na primjer električne dizalice topline s rashladnim,
apsorpcijskim uređajima te toplinskim spremnicima. Višegeneracijske jedinice imaju potencijal
uštede primarne energije i smanjenja emisije stakleničkih plinova u odnosu na konvencionalnu,
odvojenu proizvodnju korisnih oblika energije. Unatoč tome jedinice distribuirane proizvodnje
danas su najčešće spojene na elektroenergetsku mrežu kao pasivni subjekti, njihova proizvodnja
određena je ili samo pokrivanjem zahtjeva za toplinskom energijom (u tom slučaju električna
energija je nusprodukt koji se predaje u elektroenergetsku mrežu) ili iznosom poticaja koji je
moguće dobiti za električnu energiju predanu u mrežu, čime se ubrzava povrat njihove
investicije. Ovakav način rada i pristupa rijetko odgovara zahtjevima sustava, a puno češće stvara
probleme tehničke prirode. Razlozi ovakvog načina rada su višestruki, od neutvrđenih adekvatnih
regulatornih okvira, nepostojanja naprednih upravljačkih algoritama, do nedovoljno razvijene
ICT infrastrukture, kojom bi se prikupljali i slali odgovarajući procesni podatci za upravljanje
spomenutim jedinicama. Distribuirani energetski sustavi stoga će imati značajnu ulogu u
stvaranju održive energetske budućnosti.
Višegeneracijske jedinice za kombiniranu proizvodnju različitih energetskih vektora su
učinkovito rješenje za povećanje energetske učinkovitosti. Najjednostavniji koncept više
energetske proizvodne jedinice je kogeneracijska elektrana koja proizvodi toplinsku i električnu
energiju iz istog ulaznog goriva, a njene prednosti u odnosu na odvojenu konvencionalnu
proizvodnju dobro su poznate. S druge strane, električne dizalice topline smatraju se učinkovitom
i ekonomičnom alternativom postojećim kotlovima, a imaju sposobnost smanjenja emisija
stakleničkih plinova u sektoru toplinarstva (samo u slučaju kada i elektroenergetski sektor postaje
dekarboniziran).
Veća
integracija
električnih
dizalica
znatno
povećava
opterećenja
elektroenergetske mreže na koju se priključuju, pa odabir ove tehnologije neminovno zahtijeva
daljnja ulaganja u elektroenergetsku infrastrukturu.
Standardne kogeneracijske jedinice malih i srednjih instaliranih snaga (do 10 MW) imaju
zanemarivu sposobnost u reagiranju na tržišne cijene električne energije i samim time na zahtjeve
elektroenergetskog sustava. Da bi mogle reagirati na upravljačke signale, kogeneracijske jedinice
trebaju biti fleksibilnije, te je upravo iz ovog razloga prepoznata važnost povezivanja takvih
jedinica sa spremnicima topline. I dok su prednosti povezivanja kogeneracijskih jedinica sa
spremnicima topline jasne, potencijal spajanja kogeneracijskih jedinica i dizalica topline (uz
dodatak toplinskog spremnika), s ciljem povećanja fleksibilnosti u odnosu na zahtjeve sustava, je
manje istražen. Ovakav koncept imao bi dodatnu vrijednost u budućim elektroenergetskim
sustavima s velikim udjelom obnovljivih izvora energije, kad je moguće višak proizvedene
električne energije iskoristiti preusmjeravanjem između energetskih vektora. Aktualna
istraživanja sugeriraju da bi upravo kombinacija više tehnologija mogla biti optimalno rješenje u
suočavanju s promjenjivošću elektroenergetskog opterećenja, uzrokovanom povećanim udjelom
obnovljivih izvora energije, no to pitanje još nije sustavno istraženo. Nadalje, ne postoje
praktična iskustva, a vrlo je malo sustavnih modela distribuiranih višeenergetskih jedinica,
posebno onih temeljenih na kombinaciji kogeneracije i dizalica topline, te stoga njihova uloga
ostaje prilično nejasna. Neizvjestan i vrlo dinamičan kontekst budućeg energetskog sustava
otvara put istraživanju novih, složenih i inovativnih mogućnosti, temeljenih na kogeneracijskim
jedinicama i dizalicama topline, koje bi mogle biti učinkovito implementirane s ciljem
poboljšanja ekonomskih i ekoloških performansi budućih energetskih sustava.
Dodatni izazov predstavlja integracija cjelovitog višegeneracijskog sustava, koji obuhvaća cijeli
niz tehnologija prilagođenih različitim tipovima potrošača i potrošnji različitih energetskih
vektora. Složenost ovog problema povećava se ukoliko takav sustav mora biti u mogućnosti
održavati stabilnost i sigurnost opskrbe zasebno od ostatka energetskog sustava. Koncept je
posebno zanimljiv u industrijskim postrojenjima, gdje postoji relativno konstantna potrošnja više
energetskih vektora te je neprekinutost opskrbe od velike važnosti. Ovakvi integrirani koncepti
moraju imati veliku fleksibilnost u radu te biti sposobni odgovoriti na promjene u potražnji ili
cijeni energije/energenata. Ta sposobnost se očituje u mogućnosti premještanja proizvodnje s
jednog energetskog vektora na drugi kao i u skladištenju određenog energetskog vektora u
adekvatnim spremnicima, čime krajnji korisnik/potrošač uvijek dobije željenu energetsku uslugu.
Posebno je važno naglasiti da takvi sustavi ostvaruju i značajne uštede emisija CO2, a njihove
tehničke karakteristike omogućuju znatno veću integraciju neupravljivih, ali CO2 neutralnih,
obnovljivih izvora energije, kao što su solarne i vjetroelektrane.
S obzirom na to da su jedinice distribuirane proizvodnje najčešće male snage, do nekoliko MW,
nisu u mogućnosti sudjelovati na tržištu električnom energijom. Postoje izuzetci, kao što su
tržišta Njemačke i Danske, međutim s obzirom na mali utjecaj, njihovo sudjelovanje ograničeno
je na tržište dan unaprijed. Agregiranjem u jedan tržišni subjekt, poznat kao virtualna elektrana,
takve jedinice imaju mogućnost boljeg pozicioniranja na tržištu te mogućnost sudjelovanja na
više tržišta koja pružaju različite usluge.
Iz navedenih pretpostavki postavljen je cilj ovog rada: definirati cjeloviti i jedinstveni tehnoekonomski i okolišni model, te izgraditi optimizacijsku platformu za operativno planiranje
različitih višegeneracijskih sustava. Takvi sustavi moraju biti fleksibilni, kako bi mogli
odgovarati na signale i zahtjeve sustava. Koncept fleksibilnosti, promatran s dvojakog gledišta
operativne fleksibilnosti (sposobnost da reagira na cjenovne signale u stvarnom vremenu), i
fleksibilnosti planiranja (robusnost prema razvijenim scenarijima i sposobnost prilagodbe
promjenama tržišnih uvjeta), detaljno je analiziran za različite opcije. Kroz analize osjetljivosti za
svaku višegeneracijsku jedinicu pokazana je njena fleksibilnost, te izračunate ekonomske i
okolišne prednosti koje je moguće postići koordiniranjem i grupiranjem kogeneracijskih jedinica
s dizalicama topline i toplinskim spremnicima. U doktorskoj disertaciji bit će prikazana sustavna
i sveobuhvatna slika prednosti i nedostataka različitih opcija višegeneracijskih jedinica.
Ključne riječi: distribuirane višegeneracijske jedinice, dizalice topline, fleksibilnost,
kogeneracija, virtualna elektrana, više energetski sustavi, toplinski spremnici
TABLE OF CONTENTS
INTRODUCTION ......................................................................................................................................................... 1
A.
Thesis outline .................................................................................................................................................... 4
MODELLING AND OPTIMIZATION OF DISTRIBUTED MULTI-GENERATION DISTRICT HEATING
OPTIONS....................................................................................................................................................................... 6
A.
Multi-energy systems ........................................................................................................................................ 6
B.
Techno-economic and environmental modelling and optimization framework for DH DMG options............. 9
C.
B.1.
Description of DH DMG options ............................................................................................................. 9
B.2.
Unified mathematical formulation of the DMG DH modelling and optimization problem ................... 11
B.3.
Operational optimization problem formulation ...................................................................................... 13
Environmental models .................................................................................................................................... 17
C.1.
Primary energy saving ............................................................................................................................ 17
C.2.
CO2 emission reduction .......................................................................................................................... 18
C.3.
Local emission reduction ....................................................................................................................... 19
ECONOMIC ANALYSES OF DMG DH OPTIONS .................................................................................................. 20
A.
Case study description and equipment sizing ................................................................................................. 20
B.
Daily operational analysis ............................................................................................................................... 20
B.1.
Winter..................................................................................................................................................... 21
B.2.
Spring/Autumn ....................................................................................................................................... 27
B.3.
Summer .................................................................................................................................................. 31
C.
Annual operational synthesis .......................................................................................................................... 34
D.
Sensitivity analysis on base case design parameters ....................................................................................... 37
E.
Discussion of the results, planning analysis and robustness test ..................................................................... 40
ENVIRONMENTAL ANALYSES OF DMG DH OPTIONS..................................................................................... 44
A.
Primary energy saving assessment .................................................................................................................. 44
B.
CO2 emission reduction .................................................................................................................................. 45
C.
Local emission reduction ................................................................................................................................ 48
MODELLING APPROXIMATIONS .......................................................................................................................... 50
A.
Economic analyses .......................................................................................................................................... 50
B.
Environmental analyses .................................................................................................................................. 58
FLEXIBILITY - A KEY ELEMENT IN OPERATION AND PLANNING OF LOW CARBON ENERGY
SYSTEMS.................................................................................................................................................................... 61
A.
Defining flexibility.......................................................................................................................................... 61
B.
Sources of flexibility ....................................................................................................................................... 65
C.
B.1.
Storage ................................................................................................................................................... 66
B.2.
Flexible generation ................................................................................................................................. 68
B.3.
Demand response (DR) .......................................................................................................................... 68
B.4.
Electric vehicles (EV) ............................................................................................................................. 70
Quantifying flexibility .................................................................................................................................... 72
C.1.
Operational flexibility ............................................................................................................................ 72
C.2.
Planning flexibility ................................................................................................................................. 74
D.
Multi-energy system flexibility aspect ............................................................................................................ 76
E.
Operational flexibility metric of Distributed Multi-Generation ...................................................................... 80
F.
Discussion ....................................................................................................................................................... 89
MULTI-ENERGY VIRTUAL POWER PLANT ........................................................................................................ 90
A.
The concept of Virtual Power Plant ................................................................................................................ 90
B.
Modelling and economic analysis of Multi-Energy Virtual Power Plant ....................................................... 92
C.
Future work ..................................................................................................................................................... 96
CONCLUSION ............................................................................................................................................................ 98
BIBLIOGRAPHY ...................................................................................................................................................... 100
LIST OF FIGURES ................................................................................................................................................... 113
LIST OF TABLES ..................................................................................................................................................... 115
NOMENCLATURE................................................................................................................................................... 116
BIOGRAPHY ............................................................................................................................................................ 119
ŽIVOTOPIS ............................................................................................................................................................... 122
1
INTRODUCTION
There are significant efforts worldwide, from research to policy initiatives, to support the
integration of renewable energy resources into the power system. These ideas are particularly
emphasising the deployment of innovative concepts such as the Smart Grid. However, meeting
challenging environmental targets and guaranteeing secure and affordable energy to present and
future generations, require clear strategies addressing all energy sectors and not only electricity
[1]. Particularly they refer to heating, cooling and transport, as they account for over 70% in total
energy consumption, greenhouse gas emissions and pollution. In addition these sectors are, in
most cases, based on fossil primary energy and might even be harder to decarbonise than the
electricity sector.
Traditionally, different energy sectors have been decoupled from both operational and
planning viewpoints, although tight interactions have always taken place. For instance,
electricity, heat/cooling and gas networks interact in many cases through various distributed
technologies such as combined heat and power (CHP), electric heat pumps (EHP), air
conditioning devices, trigeneration of electricity heat and cooling [2], [3]. Similar interactions
between electricity and the transport sector are becoming more emphasised by means of electric
vehicles (EV) or bio-fuels and hydrogen transport, emphasizing their relevance to future planning
and operation, especially since they are becoming more accepted by end users[4].
Although today's energy systems are in fact multi-energy in nature, it is arguable how much
efficiency from coupling different energy vectors through known technologies is presently
exploited. In particular, there are very few models that are capable of assessing, in a quantitative
manner, the benefits that could be derived from different multi-generation options including
Combined Heat and Power (CHP) and Electric Heat Pumps (EHP), both considered high
efficiency technologies. In addition, particularly when associated to thermal storage, such
integrated options can bring significant benefits in terms of operational flexibility to respond to
market prices. The goal of this thesis is to develop and present a unified and comprehensive
framework and a relevant Mixed Integer Linear Programming (MILP) model suitable for
evaluating economic and environmental benefits from different Distributed Multi-Generation
(DMG) options for District Heating (DH) applications. Each option’s operational cost and benefit
and flexibility to respond to electricity market signals is analysed in detail and assessed against
2
the needed investment costs in different scenarios, with focus on the flexibility gained from
equipment integration. Detailed sensitivity analyses of different DMG configurations clearly
show what economic and environmental benefits (both at the global and the local level) can be
derived from coupling different energy vectors in current and future scenarios.
Throughout the thesis several aspects of Multi-Energy systems are researched, analyzed and
elaborated. The findings are presented in the following Chapters and the contribution of the thesis
can be summarized as:
•
Proposing a mathematical model for optimal operation of defined district heating
schemes. This has been done by: (a) defining seven different district heating types and (b)
creating a mathematical formulation that describes each of these types in a unified way.
To be more precise: the work recognizes existing district heating schemes and suggests
new, more flexible ones. All of the types are already mentioned in the literature but there
is no unified mathematical formulation describing optimal operation of all of them nor is
there a comparison of their optimal operation and benefits for choosing a certain type. The
mathematical model is formulated in such a way that each of the type can with regards to
pieces of equipment be considered as a subtype is the most comprehensive one. In spite of
its simplicity (which is indeed one of the upside of the formulation), the importance of
such a formulation is in its power to model and simulate in a synthetic way each and all
district heating type (basically covering all the fundamental schemes with market
available equipment) for any application/location. This is done in a way that relevant
decision variables are set to zero if equipment operation is not considered in the
simulation. Such a synthetic and systematic formulation for DMG system optimization is
new in the literature and allows, amongst the others, highlighting synergies (particularly
in terms of flexibility features) and differences between CHP and EHP.
•
Defining, assessing and comparing operational flexibility of different multi-generation
options. This is done in economic terms by assessing the operational benefits from
running the different types and by highlighting how the introduction of different pieces of
equipment (for instance, by coupling thermal energy storage to CHP or EHP, or by
cascading EHP to CHP) changes/improves the economic performance of the scheme
owing to the changed/improved level of flexibility and subsequent capability to respond
to market signals.
3
•
Optimal unit sizing for each defined DMG district heating type through sensitivity
analyses of both investment cost and operational cost. In particular, it has been shown
how different sizes impact the operational cost in different type solutions resulting in
determining the optimal unit sizes, taking into account the investment cost in a detailed
Net Present Value (NPV) analysis. With increasing flexibility requirements in energy
systems, presented studies thus provide key insights on the benefits of integrating
different pieces of equipment for both operational and planning perspectives.
•
Assessment of the global and local environmental benefits (global and local emission
reduction and primary energy saving) under different scenarios. The contribution is
therefore not limited only to assessment of the economic benefits of different district
heating schemes; but it also refers to the appraisal of different environmental impacts.
Amongst the others, each DMG type is analyzed for the entire period of 2011 and 2012 in
terms of CO2 emission reduction, providing knowledge of the benefits each type can have
under current conditions. In addition, different possible future scenarios are also analysed
in terms of decarbonisation of both the power grid and the fuel inputs. This enables
assessment of future environmental performance and gives insights on the environmental
robustness of given schemes under changing energy systems. Local impact analysis is
performed to complement the global one, highlighting pros and cons of different schemes
in the light of local pollution. Such systematic assessments of DMG schemes under
economic and multiple environmental criteria are new in the literature.
•
Defining flexibility of the future low carbon energy systems. The contribution of this part
is twofold: i) giving a comprehensive literature review, defining the concept of flexibility
and technical constraints which are limiting or providing additional flexibility in the
system; ii) defining operating flexibility metric for distributed multi-generation systems.
This metric evaluates additional flexibility that can be gained by shifting energy vector
production from one component to the other, depending on the signals and requests from
the upstream system. In the context of this metric, evaluation of flexibility for the most
comprehensive DMG type is given for typical seasonal days and elaborated in the context
of additional service provision.
•
Defining the concept of aggregated DMG units in Multi-Energy Virtual Power Plant.
Presented concept brings a unified framework for assessing the operational benefits as
4
well as benefits of DMG unit aggregation in a form of economic benefit matrix. The
presented analyses clearly show additional operational cost reduction from unit
aggregation, resulting from the capability to optimally select the lowest marginal cost
unit. Over a range of market price scenarios, the concept evaluates benefits of aggregating
same type units and compares it to aggregating different types of DMG units.
A. Thesis outline
First Chapter serves as an introduction to the multi-generation concept and explains the
interactions between different energy systems. Through general guidelines the idea of doctoral
thesis is explained as well as the main contributions.
Second Chapter defines different types of multi-generation units followed by a unified
mathematical formulation. The developed model enables detailed financial and environmental
analysis of each type including detailed analyses resulting from altering input parameters.
Third Chapter presents detailed financial analysis of investment and operation of distributed
multi-generation units. Through daily operational analyses, the capabilities of different types to
respond to changes in power system are defined with respect to potential savings. Optimal unit
size for each type is defined. Results are supported by a detailed investment analysis considering
the price changes in electricity, gas and discount rates which demonstrate robustness of the
proposed model.
Fourth Chapter provides a detailed environmental analysis for each multi-generation unit. All
relevant parameters, from global level perspective (CO2 emissions) to local level analyses
(primary energy savings and local emissions), are included.
Fifth Chapter presents approximations made to the mathematical model presented in previous
chapters. It elaborates, and supports this with relevant results, that approximations introduced do
not affect the quality and accuracy of the obtained results but highly improve time needed for
solving the optimization algorithm.
Sixth Chapter provides a detailed overview of flexibility concept highlighting the importance of
flexibility in future power systems operation and planning. Multi-generation unit flexibility is
defined recognizing the benefits of interaction of multiple energy vectors. The Chapter
introduced operational flexibility metric for distributed multi-generation systems at district
energy level.
5
Seventh Chapter defines the aggregation concept of multi-generation units through a concept of
multi-energy virtual power plant. Financial analyses show that, in addition to reduced costs and
emissions savings gained from optimal operation of single units, aggregation brings additional
benefits creating a better position for aggregated units at the electrical energy market.
Eighth Chapter summarizes presented work, emphasising key conclusions and scientific
contributions. It also makes suggestions and guidelines for future work.
6
MODELLING AND OPTIMIZATION OF DISTRIBUTED MULTIGENERATION DISTRICT HEATING OPTIONS
A. Multi-energy systems
In the light of the increasing interest for multi-energy systems, where multiple energy vectors
optimally interact with each other [5], district energy systems are likely to play a more and more
important role in delivering a sustainable energy future, particularly with the projected number of
people living in cities constantly increasing. In this outlook, multi-generation options for
combined production of different energy vectors [6], [7], [8] are an effective solution to increase
energy efficiency, particularly in urban areas and in district energy systems. The simplest multigeneration concept is combined heat and power (CHP), producing usable heat and electricity
from a certain input fuel. Advantages of combined heat and power plants, with respect to the
conventional separate production (SP) of electricity and heat, are well known [9], [10]. In [11]
the author elaborates on basic efficiency concepts determined by the European Directive for the
promotion of cogeneration, providing guidelines on defining primary energy savings of CHP
units which are eligible for financial benefits. Economic and environmental advantages of CHP
units for district heating (DH) systems are analyzed in [12] and [13]. Electric heat pumps (EHP),
also considered as an efficient and economic alternative to existing boilers, have a capability to
decarbonise the heating sector but are dependent on concurrent decarbonisation of electricity
generation. In addition, large integration of EHP means significant increase of electricity load in
the grid. This will eventually require further investments in the electric grid infrastructure [14].
Conventional small to medium scale CHP units are negligibly flexible in responding to electricity
market prices as they have a primary task of covering demand (of heat or in limited cases of
electricity), and in order to be able to cope with more volatile energy prices they need to be more
flexible. This is where the importance of coupling with Thermal Energy Storage (TES) is
recognized, most noticeably in Denmark where this strategy was initially supported by three level
feed-in-tariff; today even smaller distributed units are allowed to participate in the electricity dayahead market [15]. Several other countries have introduced incentives for wide spread integration
of DH CHP systems such as Germany [16]. Sweden has instead decided to focus its heating
policy on EHP systems which have already been installed in Stockholm [17]. Despite the primary
focus on EHP in Sweden, city of Linköping is also supplied by CHP coupled with TES [18]. The
7
UK is also committed to the decarbonisation of the whole energy sector and heating in particular.
In this respect, different strategies have been proposed [19], [20], [21], to follow up on
experience and successful implementations of the Danish/Swedish energy policies. Studies
researching the optimal size of CHP coupled with TES for the case of DH in the UK are
presented in [22], [23]. Today only 2% of UK consumers are connected to DH systems, but
estimates [24] suggest this technology could have a significant share in heat supply for 40 million
forecasted UK consumers in 2030.
While benefits enabled by thermal storage are fairly clear, there is less understanding of the
potential to couple CHP to EHP (and in case storage too) to gain significant flexibility to respond
to system requirements and prices. These systems could have an even greater value in future
electricity systems with large share of renewable energy sources when excess electricity can be
harvested by shifting between energy vectors [25], [26]. Some papers indeed conclude that
combination of technologies could be the optimal solution in coping with volatility of net
electricity load caused by increasing share of renewable energy sources [27], but this issue has
not been explored systematically as yet. In particular, there are neither experiences nor
comprehensive modelling framework and systematic studies on Distributed Multi-Generation
(DMG) [28], [3] options, particularly the ones based on combination of CHP and EHP and whose
role remains relatively unclear.
This uncertain but very dynamic context thus paves the way to the opportunity of exploring new
alternative and innovative DMG options based on CHP and EHP that could be effectively
deployed to increase the economic and environmental performance of future DH systems,
particularly in the UK. The potential benefits these systems bring are already recognized in
different countries [18], [29], [30], and for different district energy system applications [8], [31].
The term “equivalent cogeneration plant” is for instance introduced in [32] exemplifying the
benefits of cascading CHP-EHP through energy shifting factors. This concept can provide
significant emission reduction [32] as well as primary energy savings, as also shown in [9], [33].
The concept is further expanded to capture the flexibility benefits of multi-generation units in
[34], [35]. The idea of CHP-EHP where EHP uses cooled stored air from flue gases of CHP as a
“free fuel” is presented in [36], [37]. These papers can be considered state-of the art in terms of
coupling different units and increasing flexibility by operational shifting between different energy
8
vectors. However, there is no comprehensive market and environmental analysis of such DMG
options.
On the above premises, the aim of this work is to define a comprehensive and unified technoeconomic and environmental modelling and optimization framework for the operational and
planning evaluation of different DMG options for DH systems. The concept of flexibility, seen
here from the point of view of operational flexibility (capability to respond to price signals in
close to real time), is analysed in detail for the different options. In addition, the different DMG
options are also assessed from an investment perspective, also considering different price
scenarios, as well as from an environmental perspective considering both local (from local
pollution) and global impact (primary energy consumption and CO2 emissions) impact and in
different scenarios. In this way, a systematic and comprehensive picture of pros and cons of
different DMG options in different contexts will be provided.
The rest of the Chapter is organized as follows. Section II defines different district heating DMG
options and presents a unified MILP market based mathematical model, as well as the relevant
and consistent environmental assessment models. Numerical case study examples are elaborated
in Section III, exemplifying market driven operation of different options in a realistic UK
context. Sensitivity analyses are also run to assess the robustness of the optimal solutions, and
investment and environmental analyses further highlight the benefits gained by each DMG option
in different conditions. Conclusions and guidelines for future work are presented in Section IV.
9
B. Techno-economic and environmental modelling and optimization framework
for DH DMG options
B.1.
Description of DH DMG options
This Section introduces a comprehensive and synthetic modelling framework that, in a unified
fashion, can perform description and optimization of all the considered DH DMG options
enabling simple simulations of similar studies. For the purpose of systematic techno-economic
comparison, this thesis defines seven different DH schemes based on DMG options:
•
Type 1: A conventional boiler is considered throughout the paper as a reference case for
comparing other proposed DH schemes. A centrally connected boiler produces heat and,
through a heat network, supplies consumers. Electricity demand is satisfied through
electrical network connection and electricity is produced by the incumbent bulk
generation system and is assumed to be purchased on the wholesale market.
•
Type 2: A central EHP system provides heat using electricity from the electricity grid. As
in the first case, electricity demand is supplied through the electricity grid.
•
Type 3: Thermal energy storage (TES) is added here to the scheme proposed in type 2,
with the idea of bringing additional flexibility in the heating side.
•
Type 4: CHP-AB system. This is the prevailing DH concept with CHP units, where an
auxiliary boiler (AB) is utilized only for covering peak heat demand or for back-up.
Electricity can be bought from or sold to the electrical grid.
•
Type 5: CHP-AB-TES system. Thermal storage here provides flexibility to the CHP unit,
enabling it to decouple its operation from heat demand following.
•
Type 6: CHP-AB-EHP system. An EHP unit is cascaded to CHP. The inherent flexibility
of supplying electricity and heat from different inputs in principle significantly increases
responsiveness to price signals and variable demand.
•
Type 7: CHP-AB-EHP-TES system. This scheme combines the benefits of the previous
two types cascading CHP with a TES and EHP, gaining additional flexibility.
The relevant energy flow layout of all defined DMG types is presented in Figure 1.
10
Eimp
Eimp
Eexp
EDS
EDS
ED
ED
Faux
Haux
Boiler
ηaux
C
O
N
S
U
M
E
R
S
ED, HD
Echp
C
O
N
S
U
M
E
R
S
ED, HD
CHP
ηe, ηe
Hchp
Haux
Auxiliary boiler
ηt
d)
a)
Eimp
Eexp
Eimp
EDS
EDS
ED
ED
Ehp
Fchp
Electric Heat Pump
COP
C
O
N
S
U
M
E
R
S
ED, HD
Hhp
CHP
ηe, ηe
Echp
Hchp
Faux
Auxiliary boiler
ηt
Haux
e)
Eimp
Eexp
b)
EDS
(1-α)Echp
Eimp
EDS
ED
Ehp
Electric Heat Pump
COP
HD
TES
Hs
C
O
N
S
U
M
E
R
S
ED, HD
Hhp
TES
Hs
c)
HD
C
O
N
S
U
M
E
R
S
ED, HD
Fchp
Ehp
αEchp
CHP
ηe, ηt
Hchp
Haux
Faux
Auxiliary boiler
ηaux
f)
Electric Heat
Pump
COP
ED
Hhp
C
O
N
S
U
M
E
R
S
ED, HD
11
Eimp
Eexp
(1-α)Echp
EDS
Fchp
ED
Ehp
αEchp
CHP
ηe, ηt
Electric Heat
Pump
COP
Hhp
Hchp
TES
Hs
Haux
Faux
Auxiliary boiler
ηaux
C
O
N
S
U
M
E
R
S
ED, HD
g)
Figure 1 Energy flow layout of a) DMG type 1, b) DMG type 2, c) DMG type 3, d) DMG type 4, e) DMG type
5, f) DMG type 6, g) DMG type 7
B.2.
Unified mathematical formulation of the DMG DH modelling and optimization
problem
Mathematical models describing CHP short term operation go back to Püttgen et al. [38], [39].
An interesting approach using mixed-integer formulations describing operation of large back
steam pressure CHP units with feasible generation region can be found in [40] and [41]. In [42]
the authors present a conceptual approach to modelling each grid user in a unified way as power
nodes, consisting of its belonging efficiencies, losses, demand and generation, covering all
electro-chemical conversions. The concept of power nodes is a unified way of presenting a single
consumer/producer/storage and its interaction with the grid. However, there is no market-driven
or other type of optimization developed in the model. A complete overview on short-term
operational modelling of CHP units can be found in [43] giving an understanding on different
operational limits and technical constraints. Recent focus has been turned to smaller distributed
generation CHP units as they are becoming a relevant factor in electricity markets; for instance,
in [44] the authors present a mathematical model for µCHP unit coupled with TES, comparing
different operating scenarios for fuel cells. In [45] the authors propose a mathematical model for
economic evaluation to supply electricity, heating and cooling load to a food factory by coupling
CHP and EHP. Mehleri et al present a mathematical modelling approach to planning the entire
district energy system taking into account the investment and the operation cost of different
12
technologies [46]. However, the model relies on fixed grid prices and fixed feed-in tariffs for
photovoltaic (PV) and µCHP and thus the operation of each unit (and the entire microgrid) is
implicitly led by demand and does not exploit potential flexibility. In [47] the authors propose a
mathematical formulation for a single µCHP unit and a cluster of µCHP units, each with its own
thermal storage decoupling demand and production, enabling each unit to be more flexible and
have lower operational cost. Further flexibility in the considered microgrid is also provided by a
centralized TES unit. However, there is no consideration of EHPs. An interesting approach that
synthesizes different pieces of equipment with the goal of finding optimal superstructure of
power plant types, auxiliary equipment, and heat networks is described in [48], [49]. The concept
of superstructure is applied for the design of trigeneration system in [50]. The same group of
authors previously worked on the optimal trigeneration system design using evolutionary
algorithm component based modelling [51]. However, none of these papers elaborates on the
operational flexibility which can be gained by adding or replacing a certain piece of equipment.
Also, there is a lack of comprehensive investment studies as well as environmental impact
studies.
From the systematic literature review analysis that was carried out, it emerges that all the models
currently available are either case specific or lack methodical assessment of the marginal effect of
equipment addition/substitution. Further none can represent comprehensively and consistently
different DMG options including CHP, EHP and TES, and their flexibility in terms of operational
optimization. Additionally there is no unified approach to operational, investment and
environmental assessment. On the contrary, the developed MILP model discussed in the
following Section is capable of capturing the optimal operation of all the DMG types presented
above, namely types 2-7, and this is done by developing a synthetic problem formulation that
entails all the considered options in a unified way. Figure 2 illustrates through a flow chart the
idea of how each DMG type can be “created” by combining different pieces of equipment. More
specifically, while DMG type 2 and DMG type 4 are a single “generation” unit types, composed
only of EHP (type 2) or CHP (type 4), all remaining DMG types are structured as cascades of
either or both of these units along with storage. From the figure, it can also be appreciated how
DMG type 7 is the most complex one and all other DMG types can be considered as subtypes of
type 7.
13
DMG TYPE 7
CASCADED
EHP
TES
CHP
CASCADED
CASCADED
CASCADED
DMG TYPE 2
DMG TYPE 6
DMG TYPE 3
DMG TYPE 5
DMG TYPE 4
Figure 2 DMG types creation flow diagram
B.3.
Operational optimization problem formulation
The general operational optimization problem under analysis here can be stated as
minimize OPERATION =
T
∑ C ( t ) + C
t =1
fuel
elec _ buy
(t ) − Celec _ sell (t ) 
(2.1)
Equation (2.1) contains the objective function to minimize, that is, the operational cost for each
of DMG units taking into account fuel and electricity prices for each components of the specific
option (CHP, EHP, boiler, TES). The objective function is oriented towards assessing economic
benefits and does not include Operation and Maintenance cost (O&M) in order to better highlight
the economic performance of the different options based on energy prices. However, O&M can
be easily incorporated in the framework.
In terms of problem constraints, starting from the flow chart in Figure 2, in the following all
constraint equations for each and all DMG types are formulated in a unified and comprehensive
way by making use of a suitable superscript notation. More specifically, a superscript from 2 to 7
14
in each parameter or variable of the relevant constraint equation indicates that the relevant item
exists only in that DMG type associated to the specific superscript number.
Starting from the modelling of the CHP, in Eqs. (2.2) and (2.3), the CHP electricity output Echp
and heat output Hchp are defined at each time step starting from the fuel input Fchp through the
relevant electrical efficiency ηe and thermal efficiency ηt, respectively (superscripts in these
equations indicate the CHP can be found only in DMG types 4, 5, 6 and 7). Electrical and
thermal efficiency of the CHP unit are typically a function of the loading level. However, for the
sake of simplicity, an approximation is made in the formulation and it is assumed that efficiencies
are constant with the loading level, as for instance done in [41], [62] and [66]:
Echp ( t )
4,5,6,7
H chp ( t )
= Fchp ( t )
4,5,6,7
4,5,6,7
= Fchp ( t )
*ηe ( t )
4,5,6,7
4,5,6,7
*ηt ( t )
4,5,6,7
(2.2)
(2.3)
The CHP electricity output is limited by its upper limit Echp_max (maximum power) and by its
lower limit Echp_min (also called Minimum Stable Generation – MSG). If we indicate with Ichp a
binary decision variable that takes the value of 1 when the CHP unit is on and 0 when the CHP
unit is off, the CHP electricity production constraints can therefore be expressed as:
I chp (t ) 4,5,6,7 * Echp _ min ≤ Echp ( t )
4,5,6,7
≤ I chp (t ) 4,5,6,7 * Echp _ max
(2.4)
Since thermal power plants typically exhibit specific power variation limits (indicated here as
ramp), the power output increase or decrease between two successive time periods are subject to
the following constraint:
−ramp ≤ Echp (t ) 4,5,6,7 − Echp (t − 1) 4,5,6,7 ≤ ramp
(2.5)
The electric heat pump is present in the DMG types 2, 3, 6 and 7, in the latter two cases being
cascaded to the CHP unit. At each time step, Eq. (2.6) describes the relationship between heat
output Hhp and electricity input Ehp through the relevant Coefficient of Performance (COP) value,
while Eq. (2.7) indicates the relevant heat production constraints between the upper limit Hhp_max
and the lower limit considered here equal to zero:
15
H hp ( t )
2,3,6,7
= Ehp ( t )
0 ≤ H hp ( t )
2,3,6,7
2,3,6,7
* COP ( t )
2,3,6,7
≤ H hp _ max
(2.6)
(2.7)
The electricity Ehp used for running the EHP is generally composed of two components, namely,
an electricity flow coming from the electrical grid (Ehp_g) and, only in those cases when the EHP
is coupled to the CHP unit (that is, DMG type 6 and 7), an electricity flow coming from the CHP
(Ehp_chp). This electricity balance at the EHP input can therefore be written as:
=
Ehp ( t )
Ehp _ chp ( t )
2,3,6,7
6,7
+ Ehp _ g ( t ) 2,3,6,7
(2.8)
Superscripts in (2.8) denote how the grid is a source of electricity for the EHP in all four types 2,
3, 6 and 7 and also how in types 6 and 7 the CHP represents a further EHP electricity source.
Hence, in DMG types 6 and 7 the amount of EHP electricity input coming respectively from the
grid and the CHP unit is a result of the optimisation problem. This is highlighted by introducing
what can be called shifting factor α(t) [32] which describes the per unit fraction of electricity
produced by the CHP unit going to supply the EHP. Graphically this input-output electricity
relationship between CHP and EHP is shown in Figure 1 e) and mathematically it is formulated
as:
Ehp _ chp ( t )
6,7
= Echp ( t )
6,7
* α ( t ) 6,7
(2.9)
The general overall electricity balance is shown in Eq. (2.10), which can be viewed as the
Kirchhoff nodal law: the sum of all electricity produced locally, consumed locally, and
exchanged with the grid needs to be zero:
ED ( t )
2 ,3 ,4 ,5 ,6 ,7
(
=
1 − α (t )
6 ,7
)* E
chp
(t )
4 ,5 ,6 ,7
+ Eimp ( t )
2 ,3 ,4 ,5 ,6 ,7
− Eexp ( t )
4 ,5 ,6 ,7
− Ehp _ g ( t ) 2 ,3,6 ,7
( 2.10 )
In particular, from (2.10) it is clear how in the DMG cases 6 and 7, when the EHP is cascaded to
CHP, the fraction of electricity produced by the CHP that does not supply the EHP, namely, (1-
16
α(t)) Echp, is either consumed locally as ED or is sold to the market as Eexp. The potential
electricity amounts Eimp and Ehp_g bought from the market make up the final balance. The
electricity balance for cases 2-5 can be explained similarly.
It is interesting to point out how, similarly to the other relationships discussed in this problem
formulation but even more evidently, Eq. (2.10) actually turns into a different equation for each
different DMG type, as each of the seven types has different “consumers” (EHP and local
demand) and “producers” (CHP and the grid) of electricity. With Figure 2 in mind, this is once
again shown by superscripts 2-7: only when a superscript is associated with an existing
component for the specific DMG type under consideration can the relevant decision variable take
values different from 0.
With respect to provision of sufficient thermal energy to the customer, in each DMG type an
auxiliary boiler exists as backup and to cover peak heat demand, also providing additional
operational flexibility in the different cases. The auxiliary boiler’s operation is described by Eqs.
(2.11) and (2.12), showing the relationship between the boiler heat output Haux and the relevant
fuel thermal energy input Faux through the boiler efficiency ηaux, as well as the operational
constraints between 0 (assumed here to be the lower limit) and the boiler capacity Haux_max,
respectively:
H aux ( t )
2,3,4,5,6,7
= η aux ( t )
0 ≤ H aux ( t )
2,3,4,5,6,7
2,3,4,5,6,7
* Faux ( t )
≤ H aux _ max
2,3,4,5,6,7
(2.11)
(2.12)
Eq. (2.13) expresses the nodal heat balance in a similar way to the electrical balance in (2.10),
also capturing the inter-temporal operational characteristics of thermal storage. The total thermal
energy Hs available in the storage at each time step is in fact equal to the stored thermal energy in
the previous step plus the net heat input in the observed time step. This net input to the storage
system is represented by the sum of the heat production from CHP (Hchp), auxiliary boiler (Haux),
and heat pump (Hhp), minus the heat demand HD. For the sake of simplicity, storage losses are
neglected here (they would be very small considering the size of the system), although again they
could be easily incorporated in the model.
17
H s (t )
3,4,7
= H s ( t − 1)
3,4,7
+ H chp ( t )
4,5,6,7
+ H aux ( t )
2,3,4,5,6,7
+ H hp ( t )
2,3,6,7
− H D (t ) 2,3,4,5,6,7
(2.13)
The thermal storage capacity is defined with its hot water temperature Ts as a control variable,
with operational constraints in the storage tank between Ts_min=50°C and Ts_max=90°C. This is
described by (2.14):
Ts _ min ≤ Ts ( t )
3,4,7
≤ Ts _ max
(2.14)
The output water temperature of the electric heat pump could in principle be lower than the one
from the CHP and additional heating could be needed to reach the output temperature of the
CHP; however, here it is assumed that CO2 based EHP are used, whose output temperatures are
consistent with the CHP ones. 1
Once again, it is worth highlighting how the problem equations were all synthetically written
according to the superscript-based notation that was mentioned above, whereby specific
equations for each specific DMG type are selected from the general constraint “set” by
considering the relevant superscript. By doing so, all proposed DMG options (that encompass all
the possible CHP/EHP/TES schemes) have been presented under a unified and comprehensive
formulation.
C. Environmental models
C.1.
Primary energy saving
Combined heat and power generation can bring substantial primary energy and environmental
savings compared to separate production [9], [52]. More generally, the multi-generation energy
efficiency in terms of primary energy saving can be evaluated through static indicators comparing
the energy produced by DMG units to that produced by SP [53], [54]. If expressed with reference
to the fuel thermal input for SP, the energy efficiency indicator for CHP systems is the wellknown Primary Energy Saving or Fuel Energy Savings Ratio [55]. For a generic multi-generation
system this factor is defined as Poly-Generation Primary Energy Savings (PPES) [54]:
1
Research on CO2 heat pump heat systems suggest that the outlet temperature could have above 85°C making additional heating unnecessary
[179].
18
PPES =
F SP − F DMG
= 1−
F SP
∑
F DMG
X DMG
( X , x )∈D
(2.15)
η x SP
D is a set of different demand vectors (electricity and heat, in this case), FSP is the fuel consumed
in the separate production needed to comply with the same energy demand as in the multigeneration system. FDMG is the primary energy input to a specific multi-generation option (in
these studies all types except type one that is the reference case), including fuel for CHP and
boilers; XDMG represents the output value of each energy vector (electricity or heat). The
coefficients ηxSP represent efficiency of separate production for each energy vector. These values
are taken as reference to evaluate the primary energy necessary to produce the same amount of
energy as in a DMG plant through conventional SP means, as exemplified in Section III.F.
C.2.
CO2 emission reduction
Similarly to primary energy saving, multi-energy systems have the potential to bring CO2
emission reduction with respect to the SP means. However, this crucially depends on the overall
system carbon emissions. This issue is even more emphasised in highly fossil fuel based system
such as the current UK one. In the environmental assessment of different DMG options
conducted here, the electricity produced in a DMG unit (locally consumed or exported) is treated
as displacing electricity production from centralized power generation. The input values needed
to evaluate the CO2 emissions from each of the DMG option are the DMG fuel (gas) emission
intensity, the emission factor of the heat SP, and the Average Emission Factor (AEF) of the
electricity grid for each period of unit operation. In particular, a PCO2ER indicator can be
defined consistently with the PPES indicator as [54]:
PCO 2 ER =
F
SP
F
DMG
F
DMG
− (mCO
(mCO
( µCO
* F DMG
2)
2)
2)
=
−
1
F
SP
(mCO
∑ ( X , x )∈D ( µCOX 2 ) SP * X DMG
2)
(2.16)
x
x
Where mCO
2 is the mass of CO2 emitted to produce the useful energy output X and µCO 2 is the
emission factor associated to the production of X, expressed in (g/kWh) [54]. A detailed CO2
emission saving analysis, together with the explanations and values used for the different
emission factors, are elaborated in Section III G.
19
C.3.
Local emission reduction
While energy efficient technologies may have a lower global environmental impact in terms of
primary energy and CO2 emissions with respect to SP, local air quality could be worsened due to
the emissions of polluting by-products from burning fossil fuels in distributed cogeneration.
Typical primary pollutants that are considered in the studies are nitrogen oxides (NOx), carbon
monoxide (CO), and volatile organic compounds (VOCs – unburned, non-methane
hydrocarbons). Other pollutants, such as oxides of sulphur (SOx) and particulate matter (PM) are
mostly dependent on the fuel used and negligible if considering natural gas [55]. This issue will
be additionally emphasised in urban areas since these are typically characterized by high
concentrations of such pollutants due to traffic pollution. The local air quality regulation could
thus often be quite stringent, leading to limited acceptance of DMG systems within densely
populated areas. A systematic analysis of these issues for CHP systems have been performed in
[56], [57] and it is generalised here to the different DMG options. The local emission values
(focusing on NOx and CO for the sake of simplicity) are calculated using (2.17):
DMG
H , SP
E ,CHP
F , SP
F ,CHP
mNOx
,CO = mNOx ,CO + mNOx ,CO = µ NOx ,CO * FAB *η AB + µ NOx ,CO * Fchp *η e
(2.17)
This model is relevant to the local emission balance approach discussed in [56], where the
emission impact from central generation electricity (both consumed and displaced) is not
accounted for since it is considered “far enough” from where the receptors in urban areas are.
20
ECONOMIC ANALYSES OF DMG DH OPTIONS
A. Case study description and equipment sizing
Case study applications have been performed to illustrate and test the model developed and to
compare the different options. For the sake of example, unit sizes are optimized to supply about
1000 consumers of different consumption profiles [58], although the model is general so as to
cover any application. Equipment sizing is a result of sensitivity analyses performed for each type
and will be discussed below. However, the results for the base case are already presented in Table
1 for easier understanding of the following discussions. For the EHP, a CO2-based ground-source
or water-source heat pump is taken as a reference for centralised applications [59], while for the
CHP system a conventional gas-fired internal combustion engine is used. Although even for a
same technology different size CHPs could have different performance, same efficiencies have
been assumed here for the different systems in order to focus on the comparison in terms of
sizing and flexibility for the various options. For all DMG types, the presence of a 6 MW backup
boiler is assumed but it is not considered in the comparison. For the investment and operation
analyses CHP sizes of up to about 5 MWt were considered. The same scale was used for EHP.
Table 1 Optimal sizes and relevant efficiencies of units for different DMG DG schemes
CHP (kWe)
TYPE 1
TYPE 2
TYPE 3
TYPE 4
TYPE 5
TYPE 6
TYPE 7
-
-
-
4000
3000
2500
2000
Electrical
efficiency
0.35
0.35
0.35
0.35
Thermal efficiency
0.45
0.45
0.45
0.45
Boiler (kWt)
6000
6000
6000
6000
6000
6000
6000
Boiler efficiency
0.85
0.85
0.85
0.85
0.85
0.85
0.85
TES (m3)
-
-
200
-
200
-
200
EHP (kWt)
-
5000
3000
-
-
2000
2000
COP
-
3.0
3.0
-
-
3.0
3.0
B. Daily operational analysis
Operational simulations for each DMG option have been conducted for typical winter, spring,
summer, and autumn days to describe representative performance throughout the year. Operation
is optimized based on available day-ahead market hourly electricity prices and for given fuel
prices (natural gas is assumed here to be input fuel for CHP units). The electricity prices are
21
taken from [60] and electricity and heat profiles from [58]. Gas prices are an average of recent
gas prices for power producers in UK taken from [61]. In this paper, for the sake of simplicity
perfect forecast of loads and knowledge of energy prices are assumed.
B.1.
Winter
In Figure 3 electricity and heat profiles for a winter day are shown together with the
7000
0,12
6000
0,10
0,08
4000
0,06
3000
0,04
2000
1000
0,02
0
0,00
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Energy Demand (kW)
5000
Electricity prices (€/kWh)
corresponding electricity prices.
ED
HD
Buy EE price
Figure 3 Demand profiles and electricity prices for a specific winter day
Figure 4. shows the operation of the CHP-AB (DMG type 4) option. The auxiliary boiler is
sized so as to cover 20% of annual heat consumption as suggested in [52], [62]. It should be
noticed that these numbers come from practical experiences from installing DH systems. When
compared to [37], [63] it can be noticed that some previous research did not rely on these
experiences which sometimes resulted in oversizing units. These CHP schemes have very limited
flexibility as already reported in [64]; they behave as large boilers following heat demand and
selling excess electricity (as a “by-product”) to the grid.
22
4500
4000
Electricity (kW)
3500
3000
2500
2000
1500
1000
500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time(h)
Eimp (kW)
Eexp (kW)
Echp (kW)
ED (kW)
7000
6000
Heat (kW)
5000
4000
3000
2000
1000
-1000
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
Haux (kW)
Hchp (kW)
HD (kW)
Figure 4 Daily operation of DMG type 4 for a typical winter day – electricity and heat
Coupling thermal storage to CHP and AB units creates flexibility (DMG type 5). In this case, the
CHP unit is not solely driven by the consumers’ heat demand, as it is able to store excess heat
during high electricity prices and thus sell electricity at favourable market prices. On the other
23
hand, during periods of low electricity prices the CHP can be offline and the heat demand can be
covered by the AB and TES. Coupling cogeneration and thermal storage has therefore the
potential to provide demand response since flexibility arises from shifting between different
energy vectors responding to real-time prices without reducing consumers’ comfort [65]. Figure 5
presents the daily operation of this scheme for a typical winter day. The mentioned flexibility can
be seen in the early morning hours (3:00-5:00) where the spark spread between electricity and
gas (whose working definition will be here the difference between the electricity price and the
gas price adjusted by the electrical efficiency) is unfavourable for CHP to be operational.
2500
1500
1000
500
0
-500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Electricity (kW)
2000
Time (h)
Eimp (kW)
Eexp (kW)
Echp (kW)
ED (kW)
24
6000
5000
4000
Heat (kW)
3000
2000
1000
-1000
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
-2000
-3000
Time (h)
STOR (charge/discharge)
Hchp (kW)
Haux (kW)
HD
Figure 5 Daily operation of DMG type 5 for a typical winter day – electricity and heat
In this period, the entire heat demand is covered by discharging the thermal storage, while
electricity is bought from the market. On the other hand, in the periods 12:00-13:00 when
electricity prices are high, the CHP operates at its maximum storing excess heat in the TES.
Cascading CHP and EHP creates virtual competition between these two units based on the hourly
spark spread. The heat pump can be supplied either from the CHP or from the grid, while CHP
production is not strictly dictated by heat demand, thus leading to more profit-oriented operation.
This competition can be clearly seen during the early morning periods, between 1:00-5:00, in
Figure 6 In these periods the heat pump uses electricity produced by CHP instead of grid
electricity minimizing the overall operational cost. An interesting sag in the exported (sold)
electricity can be noticed at the period 8:00-9:00 when, despite high electricity price on the
market, the heat peak demand drives the utilization of the heat pump. In general, in fact, high
market prices of import electricity mean it is more feasible for the EHP to use electricity
produced by CHP to cover the consumers’ heat demand. However, although the flexibility in this
DMG configuration is increased relative to previous cases, during peak heat demand the DMG
system still tends to be driven by demand rather than by market price signals. This suggests that
electricity-heat coupling is only partially buffered in this scheme (at least at peak consumption
times).
25
3500
Electricity (kW)
3000
2500
2000
1500
1000
500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
Eimp (kW)
Eexp (kW)
Echp (kW)
ED (kW)
7000
6000
Heat (kW)
5000
4000
3000
2000
1000
-1000
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
Hhp (kW)
Haux (kW)
Hchp (kW)
HD (kW)
Figure 6 Daily operation of DMG type 6 for a typical winter day – electricity and heat
Following up on the conclusions of the previous paragraph, it can be appreciated that addition of
thermal storage to the CHP-EHP cascade creates a more flexible DMG unit primarily driven by
market signals (DMG type 7). Now both units optimize their operation based on the electricity
prices but also based on the state of charge of TES. This can be seen during the periods 7:00 –
26
10:00 in Figure 7. High electricity prices push the CHP to operate at its maximum. However, the
heat produced by CHP is not sufficient to cover the entire heat demand. Comparing the same
periods for operations of DMG type 6 and 7 it can be noticed that in the mentioned period the
EHP is used to cover the remaining heat demand. Since the price of electricity is high during this
period, in DMG type 7 the TES is instead discharged and thus lower overall operational costs are
achieved. Comparing the periods between 11:30 and 12.30 for DMG type 6 and DMG type 7, the
benefits from adding TES become apparent. Addition of thermal storage enables the CHP system
to produce maximum output exploiting the benefits of high market prices, since the TES enables
it to store excess produce heat. In contrast, in DMG 6 (without TES) CHP and EHP are following
the heat demand having the EHP being utilized almost at its maximum. The flexibility benefits
obtained in these periods reflect in operational savings of the last DMG option.
2500
1500
1000
500
0
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Heat (kW)
2000
-500
Time (h9
Eimp (kW)
Eexp (kW)
Echp (kW)
ED (kW)
27
6000
5000
4000
2000
1000
0
-1000
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Heat (kW)
3000
-2000
-3000
-4000
Stor. (charge/disch)
Time (h)
Hchp (kW)
Haux (kW)
Hhp (kW)
HD
Figure 7 Daily operation of DMG type 7 for a typical winter day – electricity and heat
B.2.
Spring/Autumn
Spring and autumn periods are characterized by lower ratio of heat and electricity demand. This
is shown in Figure 8. Some behaviour patterns in terms of consumption can be still readily
recognized (morning and evening peak demands, in particular).
DMG types 4 and 5 show a very similar behaviour as for the winter day. DMG type 4 still
operates with negligible flexibility, similar as a boiler covering heat demand. Adding TES
improves the flexibility as the operation of the DMG type 5 is not constrained only by heat
demand. Thermal storage enables storing excess heat in times of high electricity prices and also
lowering CHP production during periods of low electricity prices. As the contribution of this
work is primarily in showing flexibility benefits from coupling units, the focus will be on DMG
28
types 6 and 7. Hence, the behaviour of DMG type 6 for the specific spring/autumn day is shown
3500
3000
Heat (kW)
2500
2000
1500
1000
500
-500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
Hchp (kW)
in
Haux (kW)
Hhp (kW)
HD (kW)
0,14
3000
0,12
2500
0,10
2000
0,08
1500
0,06
1000
0,04
500
0,02
0
0,00
ED
HD
Buy EE price
Figure 8 Demand profiles and electricity prices for a specific spring/autumn day
Electricity price (€/kWh)
3500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Energy demand (kW)
Figure 9.
29
3000
Electricity (kW)
2500
2000
1500
1000
500
-500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
Echp (kW)
Eexp (kW)
Eimp (kW)
ED (kW)
3500
3000
Heat (kW)
2500
2000
1500
1000
500
-500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
Hchp (kW)
Haux (kW)
Hhp (kW)
HD (kW)
Figure 9 Daily operation of DMG type 6 for a typical sprig/autumn day – electricity and heat
For the considered day the heat demand is, for the most part, covered by the CHP unit with the
exception of morning hours. Again the virtual competition between CHP and EHP, observed in
the simulations for the winter period, can be observed during periods of large difference between
sell and buy electricity prices (periods between 12:00-13:00 and 21:00-22:00). During morning
30
peak demand, namely, 6:00-7:00, the cooperation of these two units is manifested through using
the electricity produced in the CHP for running the EHP.
Operational behaviour for DMG type 7 is shown in Figure 10. At the beginning of the simulation
period, 0:00-1:00, the difference between sell and buy price is negligible and it is feasible to
cover the heat demand from the EHP. Same behaviour can also be observed during hours 4:005:00. Between those two periods the electricity price difference is high while the heat demand is
quite low; this favours storage discharging as it is not feasible to start the operation of the CHP
and producing heat from the EHP is too expensive. During the observed period (1:00-4:00), CHP
would be operating only in heat following mode selling produced electricity into the grid at, very
likely, unfavourable prices. In periods of peak demand (8:00-10:00), the heat demand is covered
by the CHP. In case there is excess demand of heat, this is covered by discharging the thermal
storage (period 8:00-9:00). Hence, the advantage of adopting DMG type 7 can be seen with
respect to not only the spark spread arbitrage of gas and electricity prices as with type 6, but also
through the capability to respond to the difference between sell and buy electricity prices by
exploiting storage charge/discharge as this difference increases.
2500
1500
1000
500
0
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Electricity (kW)
2000
-500
Time (h)
Echp (kW)
Eexp (kW)
Eimp (kW)
ED (kW)
31
3500
3000
2500
Heat (kW)
2000
1500
1000
500
-500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
-1000
-1500
-2000
Time (h)
Hs (kW)
Hchp (kW)
Haux (kW)
Hhp (kW)
HD (kW)
Figure 10 Daily operation of DMG type 7 for a typical sprig/autumn day – electricity and heat
B.3.
Summer
During summer the ratio between peak electricity and heat consumption is reduced to around
1:1.4. Figure 11 shows electricity prices and electricity and heat demand for the considered
1400
0,16
1200
0,14
0,10
800
0,08
600
0,06
400
0,04
200
0,02
0
0,00
ED
HD
Buy EE Price
Electricity prices (€/kWh)
0,12
1000
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Energy demand (kW)
summer day.
32
Figure 11 Demand profiles and electricity prices for a specific summer day
The results for DMG type 4 and type 5 are again not shown and can, for the most part, be
understood from simulations shown for a typical winter day. More specifically, for the summer
day operation of DMG type 4 the entire heat demand is covered by the AB since the trade-off
between producing in CHP and selling excess electricity to the grid is less favourable than
running the boiler and purchasing electricity from the market.
Figure 12 shows the operation of the CHP coupled with EHP (DMG unit 6).
1200
800
600
400
200
0
-200
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Electricity (kW)
1000
Time (h)
Echp (kW)
Eexp (kW)
Eimp (kW)
ED (kW)
33
1400
1200
Heat (kW)
1000
800
600
400
200
-200
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
Hchp (kW)
Haux (kW)
Hhp (kW)
HD (kW)
Figure 12 Daily operation of DMG type 6 for a typical summer day – electricity and heat
Unlike in the spring and winter cases, during summer most of the heat demand is supplied from
EHP which is driven by electricity from the grid. This can be explained with the minimal stable
generation (MSG) constraint of the CHP. For the larger period of the day, starting the CHP would
mean excess heat production which cannot be stored. Adding TES, in DMG type 7, thus enables
more flexible CHP operation, especially during peak electricity prices. This is shown in Figure 13
(periods 9:00-11:00 and 18:00-19:00).
34
2500
Electricity (kW)
2000
1500
1000
500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
-500
Time (h)
Echp (kW)
Eexp (kW)
Eimp (kW)
ED (kW)
2500
2000
Heat (kW)
1500
1000
500
-500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
-1000
-1500
Hs (kW)
Time (h)
Hchp (kW)
Haux (kW)
Hhp (kW)
HD (kW)
Figure 13 Daily operation of DMG type 7 for a typical summer day – electricity and heat
C. Annual operational synthesis
Analyzing daily operation profiles provides insight into understanding market behaviour of
each DMG type and, for the presented studies, interactions between different units and the arising
flexibility. Market conditions and load demand correlations change throughout the year so the
35
real value of flexible systems needs to be obtained by performing annual DMG type operation
analysis. This optimal operation simulates the behaviour of each DMG type for different demand
profiles, different market prices, and therefore different operational aspects of each unit type.
Hence, the same simulations as for daily operation are conducted taking the day-ahead UK
electricity market prices for the whole 2011. The flexibility of different DMG types,
demonstrated in the previous section, is supported by annual operational cost savings achieved by
unit coupling. Table 2 provides an incremental market analysis of the benefits of the different
DMG-based district heating options.
Table 2 Annual operational costs of different DMG types
UNIT TYPE
OPERATIONAL COST (€/a)
Δ TO REFERENCE CASE (%)
TYPE 1
625,440
-
TYPE 2
479,599
-23.3
TYPE 3
445,660
-28.7
TYPE 4
462,691
-26.0
TYPE 5
422,436
-32.5
TYPE 6
342,684
-45.2
TYPE 7
312,198
-50.1
As said earlier, boilers have no flexibility and there is no room for operational optimization.
The heat is produced when needed and electricity purchased to cover electricity demand. The
annual operational cost of this unit is taken as a reference value when determining flexibility and
operational savings that can be achieved from alternative options.
As also mentioned above, large and decentralized electric heat pumps are becoming a more
interesting option as they have a significant potential to manage variability in renewable energy
sources, especially wind, production [66]. Coupling EHP with thermal storage unit creates
flexibility in terms of delaying or pushing operation of EHP not strictly related to heat
consumption. Comparing these two options (types 2 and 3) to the reference case annual operation
savings of 23.3% and 28.7% can be achieved.
While CHP systems may be very efficient, the flexibility of schemes based on gas turbines or
combustion engines are limited, given the sole role of covering either heat or electricity demand.
Adding storage to such units provides them with flexibility to respond to favourable electricity
market prices. The results presented confirm this option has significant advantage compared to
the base boiler case, as it achieves 32.5% annual operation cost reduction. The possibility of
36
optimizing both sale and purchase of electricity also bring benefits compared to the EHP only
options.
Techno-economic evaluation of a CHP-EHP cascade has gained little attention in the past,
despite of the potential energy advantages this approach could bring. In fact, the annual operation
analysis of such a concept supports the conclusions based on daily simulations as cost savings
compared to the reference case are 45.2% (no storage) and 50% (with storage). If DMG options 6
and 7 are compared to CHP-AB or CHP-AB-TES units, the savings are somewhat lower although
substantial, namely, 25.9% and 32.5% respectively compared to CHP-AB and 18.9% and 26.1%
respectively when compared to CHP-AB-TES. The additional benefits from CHP-EHP coupling
are thus very significant in general. In addition, each unit option has been assessed according to a
day-ahead market analysis only. However, high flexibility, as the one shown for DMGs type 6
and 7, suggests these plants could increase their revenue even further by entering multiple
markets, e.g. intraday balancing markets and ancillary services (reserve) markets, which is object
of ongoing research.
37
D. Sensitivity analysis on base case design parameters
As already mentioned in Section II, sizing of the units is a result of sensitivity analyses
conducted for different unit sizes and DMG market driven operation though the whole year. As
DMG types 1-4 are more or less heat demand led, unit sizing in those types is essentially dictated
by peak heat demand and there is little flexibility for optimal sizing. DMG unit types where EHP
and/or TES are coupled to CHP have instead been analyzed in detail. The investment cost for
units are taken from [20], [23]. Investment lifetime of 15 years and discount rate of 5% were
initially considered. The costs and factors taken for the investment analyses are shown in Table 3.
Table 3 Input data for the investment analyses
CHP (€/kWe)
600
EHP (€/kWt)
240
3
840
TES (€/m )
Discount rate (%)
5
Investment period (years)
15
The results of investment analyses are presented in Figures 14-16.
430000
Operational cost (€/a)
425000
420000
415000
410000
405000
150
200
250
TES size (m3)
300
350
400
38
160000
140000
120000
80000
60000
40000
20000
0
150
200
250
300
350
400
TES SIZE (m3)
Figure 14 Results for sensitivity analyses for DMG type 5
380000
Operational cost (€/a)
NPV (€)
100000
370000
360000
350000
340000
330000
320000
310000
300000
290000
2500
3000
3500
1000
1500
4000
2000
EHP size (kWt)
2500
CHP size (kWe)
39
NPV (€)
1000000
900000
800000
700000
600000
500000
400000
300000
200000
100000
0
-100000
2500
3000
3500
1000
1500
CHP size(kWe)
4000
2000
2500
EHP size (kWt)
Figure 15 Results for sensitivity analyses for DMG type 6
Operational cost (€/a)
360000
350000
340000
330000
320000
310000
300000
300
250
290000
280000
200
270000
1000
1500
EHP size (kWt)
150
2000
2500
TES size
(m3)
40
1600000
1400000
NPV (€)
1200000
1000000
800000
600000
400000
300
200000
250
200
0
1000
1500
150
2000
TES size
(m3)
2500
EHP size (kWt)
Figure 16 Results for sensitivity analyses for DMG type 7
E. Discussion of the results, planning analysis and robustness test
Investment analyses support the results shown for daily and annual unit operation. For DMG type
5, the flexibility of the unit is dictated by the size of TES coupled to CHP. On the other hand,
larger TES means higher investment cost, which is not paid back by benefits compared to the
base case of boiler-produced heat. Therefore optimal size of the TES is a trade-off between
benefits and investment cost.
For DMG types 6 and 7, the electrical size ratio of CHP and EHP, in terms of heat size of the
EHP given the efficiency assumptions, has higher impact on operational cost than the size of
TES. The net present value (NPV) analyses confirm the operational cost result; the thermal ratio
of CHP to EHP size close to 1 is the optimal unit selection.
The NPV studies are based on cash flow analyses assuming steady state prices over the plant
lifetime. This assumption is neglecting investment factors such as price changes and regulatory
factors, but the results are nevertheless an extremely valuable indicator of benefits that can be
gained through different investment decisions. In addition, in order to further strengthen the
investment analysis results, a robustness test was performed. There are multiple techniques to
41
cope with long term uncertainties that effect investment decisions, such as multi stage uncertainty
analysis [67], [68], which are outside the scope of this paper, the discount rate value was initially
modified in order to show how investments depend on initial assumptions [69]. In Figure 17
sensitivity analysis with respect to different discount rate, namely, 3%, 5%, 7% and 10%, is thus
shown.
2500000
2000000
NPV (€)
1500000
1000000
500000
0
3%
5%
7%
10%
-500000
Discount rate
-1000000
TYPE 5
TYPE 6
TYPE 7
Figure 17 NPV dependency on discount rate for different DMG types
It can be easily noticed that DMG types 6 and 7 have a positive NPV value even with higher
discount rate values (when benefits gained from operational savings are somehow weighted less
relative to higher initial costs), confirming that the intrinsic flexibility to respond to price signals
also brings longer term benefits in terms of investment planning even for demanding investment
return requirements.
As a further proof of planning option robustness, sensitivity analyses with variation of gas and
electricity prices have been conducted. Input for these analyses were the optimal unit sizes
calculated in the previous section and this was taken as the base scenario (Scenario 1). Gas and
electricity prices were then varied ±50% over 8 scenarios. The scenarios presented in Figure 18
are as follows:
1) Scenario 1: Gas and electricity prices are the same as today and these results have already
been presented in Table II;
41
42
2) Scenario 2: Gas prices are increased by 50% compared to Scenario 1; electricity prices are
the same as in Scenario 1;
3) Scenario 3: Electricity prices are increased by 50% compared to Scenario 1; gas prices are
the same as in Scenario 1;
4) Scenario 4: Both electricity and gas prices are increased by 50% compared to Scenario 1;
5) Scenario 5: Gas prices are decreased by 50% compared to Scenario 1; electricity prices
are the same as in Scenario 1;
6) Scenario 6: Electricity prices are decreased by 50% compared to Scenario 1; gas prices
are the same as in Scenario 1;
7) Scenario 7: Both electricity and gas prices are decreased by 50% compared to Scenario 1;
8) Scenario 8: Electricity prices are increased by 50% compared to Scenario 1; gas prices are
decreased by 50% compared to Scenario 1;
9) Scenario 9: Gas prices are increased by 50% compared to Scenario 1; electricity prices are
decreased by 50% compared to Scenario 1;
160%
140%
120%
100%
80%
60%
40%
20%
DMG 5 savings
DMG 6 savings
Scenario 9
Scenario 8
Scenario 7
Scenario 6
Scenario 5
Scenario 4
Scenario 3
-40%
Scenario 2
-20%
Scenario 1
0%
DMG 7 savings
Figure 18 Sensitivity analyses for DMG types 5, 6 and 7
The results presented in Figure 18 show the dependency of the compared results on the unit
sizing that was performed for base case (in Section III D). This particularly refers to the size of
the CHP as this is relevant to the electricity that can be produced by the CHP and sold back to the
market. DMG type 5 CHP coupled with TES shows stronger dependence on gas and electricity
42
43
price ratio than other two types. Larger optimal size of the CHP unit results in the operation that
is highly oriented towards selling excess electricity in order to reduce operational cost. For this
reason, scenarios where price difference of electricity to gas increases in favour of electricity
(Scenario 5 and Scenario 8), results in higher operational savings for DMG type 5.
For different gas and electricity prices market driven analyses of DMG types 6 and 7 the results
shown in Figure 18 support the conclusions from the previous Section. In fact, they even suggest
that current gas and electricity price spark spread is one of the least favourable in supporting the
economic benefits from these flexible units. In scenario 8, where gas prices are increased by 50%
and electricity prices are reduced by 50%, district systems based on boiler are even economically
favourable to this DMG option. This sensitivity to prices of DMG type 5 as well as different CHP
size compared to DMG type 7 is also shown when comparing these two types.
43
44
ENVIRONMENTAL ANALYSES OF DMG DH OPTIONS
A. Primary energy saving assessment
For each of the DMG options that have been analysed, the PPES indicator has been used to
assess the relevant primary energy saving with respect to type one (heat produced in a district
heating boiler and electricity produced in centralized generation) which is considered the SP
reference (and is assigned a PPES value 0.0).
Primary energy saving is by definition independent of the fuel (see for instance references [53],
[54], where a comprehensive analysis of the differences between primary energy and emissions
and of their relation with fuels is given). However, the reference efficiencies could be selected
according to different arguments, for instance by considering only those power plants with the
same type of fuel as the considered DMG system as opposed to the whole power system. Both
types of studies were therefore run. More specifically, in order to compare the different DMG
options considered here with the overall power system, the average power system efficiency of
the UK in 2012 was considered, estimated to be around 0.38 [70], also taking into account
transmission and distribution losses in the order of 7%. Reference boiler efficiency equal to 0.85
was also assumed. In addition, the PPES indicator was also calculated with respect to state of the
art gas combined Cycle Gas Turbines (CCGT) with 55% efficiency and gas boilers with 95%
efficiencies. The relevant results are shown in Table 4. The results clearly show that the highest
primary energy savings are achieved in DMG type 7, where a PPES in the order of 40% can be
achieved with respect to the average power systems production. In addition, while the PPES is
lower when considering only gas CCGT units in the comparison, still DMG options provide
substantial energy saving of up to 31% for DMG type 7. It is also very interesting to highlight
how, from the comparison between CHP-EHP mixed schemes (type 6 and 7) and CHP only
(types 4 and 5), the higher flexibility DH schemes 6 and 7 bring substantial benefits with respect
to “conventional” simultaneous production of energy vectors in CHP. Furthermore, moving from
type 5 to type 7, the additional energy savings are much more substantial when CCGT plants are
used as the reference case (with benefits that pass from 6.5% to 31%) than when the average
power system is considered. This is because the combination of CHP and EHP allows an optimal
use of local resources and at higher efficiency than with CHP or EHP only. This also implies that
44
45
less electricity is exchanged externally, which is particularly beneficial when the reference
system becomes more efficient as in the case of CCGT plants.
Table 4 Multi-generation primary energy savings for different DMG options
PPES (UK avg. 2012)
PPES (CCGT)
DMG TYPE 1
0.000
0.000
DMG TYPE 2
0.181
0.309
DMG TYPE 3
0.183
0.318
DMG TYPE 4
0.268
0.065
DMG TYPE 5
0.286
0.079
DMG TYPE 6
0.386
0.284
DMG TYPE 7
0.403
0.310
B. CO2 emission reduction
Regarding the CO2 emission reduction assessment of the different DMG options, a gas emission
factor value of 0.19 kg CO2/kWh was used for the CHP units and the boiler, as shown in [71].
Emission rate for electricity values are reconstructed as in [72] using the electricity generation
mix for the entire 2011 and 2012 [60] and the standard emission rates for each fuel type [73]. A
transmission and distribution loss factor of 1.1 (that is, 10% losses) was taken into account too.
The resulting grid half-hourly AEF for the years 2011 and 2012 are shown in Figure 19.
45
46
800
700
CO2 emission (kg/MWh)
600
500
400
300
200
100
0
1
1441
2881
4321
5761
7201
8641 10081 11521 12961 14401 15841 17281
Settlement period
2011
2012
CASE A (2011 reduced 20%)
CASE B (2011 reduced 30%)
Figure 19 Half hourly Average Emission Factor for UK in 2011 and 2012
For a comprehensive environmental analysis of the different options, several case studies were
defined:
1) Case 2011: AEF as in 2011 for each half hour period as shown in Figure 18, gas emission
factor of 0.19 kg CO2/kWh;
2) Case 2012: AEF as in 2012 for each half hour period as shown in Figure 18, gas emission
factor of 0.19 kg CO2/kWh;
3) Case A: AEF reduced by 20% compared to the one in 2011, gas emission factor of 0.19
kg CO2/kWh;
4) Case B: AEF reduced by 30% compared to the one in 2011, gas emission factor of 0.19
kg CO2/kWh;
5) Case A1: AEF reduced by 20% compared to the one in 2011, gas emission factor of 0.17
kg CO2/kWh;
6) Case B1: AEF reduced by 30% compared to the one in 2011, gas emission factor of 0.17
kg CO2/kWh.
46
47
Case A and Case B, corresponding to reduced grid intensity scenarios, represent potential
emission reduction that can be achieved at the system level by for instance integrating Carbon
Capture Storage to coal/gas fired power plants and by increasing penetration of wind. An
approximation has been made assuming that system dynamics and flexibility as well as emission
profile would not change dramatically. More correct values could be gained by simulating future
power generation mix and its operation according to forecasted fuel and electricity prices, but this
is out of the scope of this paper. The approximations made should be viewed in the light of
bringing a high level view on the DMG environmental impact in future less carbon intensive
power systems and as such the studies are valuable in terms of strategic understanding.
Case A1 and Case B1 are based on the ongoing research, conducted mostly by the gas companies,
on mixing biogas or bio methane (or other types of low-carbon gas such as hydrogen produced
through renewables) with natural gas [74], [75]. In this respect National Grid published a
research report on energy and economic potential of renewable gas [76], while the Department of
Energy and Climate Change issued a guide for producers to connect to the existing gas grid [77].
Bio-methane producers in UK are even, under Renewable Heat Incentive, eligible for an
incentive for each kWh of bio methane injected into the gas grid [78]. The presented cases
assume for illustrative and exemplificative purposes that injecting low-carbon gas into the gas
grid will reduce the overall gas emission factor by 10%.
The results of the six scenario studies are presented in Figure 20.
5000000
CO2 emissions (kg/a)
4500000
4000000
3500000
3000000
2500000
2000000
1500000
1000000
500000
0
DMG TYPE 1 DMG TYPE 2 DMG TYPE 3 DMG TYPE 4 DMG TYPE 5 DMG TYPE 6 DMG TYPE 7
2011
2012
CASE A
CASE B
CASE A1
CASE B1
Figure 20 DMG options carbon footprint
47
48
The results presented in Figure 20 confirm the previous findings: CO2 emissions of specific
DMG unit are highly dependent on the system level emissions. As the UK electricity production
is based on fossil fuels, and thus has high average CO2 system emission rate, DMG types 2 and 3
based on EHP result in highest annual emissions. Still, these two options are environmentally
more favourable than boiler based DH options. Emissions are further reduced for DMG types
based on CHP; in some cases over 50% reduction can be achieved. Scenario emissions for 2011
and 2012 suggest that DMG types 4 and 5 are environmentally friendlier than more flexible
DMG units type 6 and 7. As the latter two (type 6 and 7) are smaller in size, by 33%, they
produce less electricity and therefore displace less electricity that would need to be produced in
highly polluting central power plants, in line with the above discussions on the primary energy
saving. This can be easily seen by comparing scenarios for 2011 and 2012; DMG types 4 and 5
are the only ones with lower CO2 emissions in the more carbon intensive 2012. Analyses for
cases with lower system CO2 emissions (Cases A, B, A1, B1) suggest that even just 10%
decrease in gas carbon intensity could retain substantial environmental benefits for gas based
DMG technologies. These results are relevant to ongoing discussion on the role that gas could
play in the next years, as mentioned above. All these options could be readily evaluated under the
presented unified modelling framework.
In addition, although not captured in the studies performed here, flexible DMG options can also
bring benefits in terms of capability to balance intermittent renewable resources in close to real
time and then integrate additional renewable at a system level. These further benefits are being
analysed systematically by the authors in ongoing research.
C. Local emission reduction
For the local emission assessment, NOx and CO emissions have been considered here as they are
the biggest concern from a regulatory point of view. The results for the different DMG options
are reported in Table 5. For this purpose, the following emission factors have been used: µ𝑊
𝑁𝑂𝑥 =
𝐻,𝑆𝑃
𝐻,𝑆𝑃
700 mg/kWhe and µ𝑊
𝐶𝑂 = 400 mg/kWhe for CHP units [79] and µ𝑁𝑂𝑥 = 200 mg/kWht and µ𝐶𝑂 =
40 mg/kWht. for boilers [56].
48
49
Table 5 Local emissions assessments for different DMG options
DMG
type 1
DMG
type 2
DMG
type 3
DMG
type 4
DMG
type 5
DMG
type 6
DMG
type 7
NOx (t/a)
3.151
0.196
0.117
6.932
7.555
3.34
3.150
CO (t/a)
0.630
0.039
0.023
3.607
4.097
1.86
1.791
District heating units based on EHP emit the smallest amount of local emissions as they only use
fossil fuels to run auxiliary boiler for backup when the price spark spread between bought
electricity and gas is in favour of gas. These results were expected and the idea of this analysis
was to evaluate the best option from the CHP based ones (DMG types 4, 5, 6 and 7) and to
compare them with base case of boiler DH units. The results support the results from the studies
performed above; besides being the best option economically, DMG type 7 has the lowest
environmental impact. When comparing this option to the reference one, where electricity is
produced in conventional power plants and local emissions are thus displaced, DMG type 7 and
boiler based DH options have the same emissions in terms of NOx. As the impact of these
pollutants is relatively limited to in the order of tens of kilometres, these analyses are valuable
when deciding on future investments, especially in high populated areas where recipients are
close to the pollutant emitters.
49
50
MODELLING APPROXIMATIONS
As mentioned earlier, efficiencies have been modelled as constant values; this is common
approximation of the real values made to speed up the calculations. As the results might be
influenced by these approximations, this section extends the mathematical formulation in order to
capture the non linear behaviour of the efficiencies with respect to the operations points of the
output vectors, heat and electricity. Equations (5.1) and (5.2) have been changed to:
=
Echp ( t )
Fchp (t ) 4,5,6,7 *ηe1 (t ) 4,5,6,7 + I chp (t ) 4,5,6,7 *ηe 2 (t ) 4,5,6,7
(5.1)
=
H chp ( t )
Fchp (t ) 4,5,6,7 *ηt1 (t ) 4,5,6,7 + I chp (t ) 4,5,6,7 *ηt 2 (t ) 4,5,6,7
(5.2)
4,5,6,7
4,5,6,7
Efficiencies ηe and ηt are modelled as piece-wise linear functions which enabled maintaining the
linearity of the entire optimization problem. Although CHP units of different sizes could have
different nominal performance, same full-load efficiencies have been assumed here for different
sizes in order to focus on the comparison in terms of sizing and flexibility for the various DMG
options. On the other hand, in order to preserve the same CHP efficiency curve shape with
respect to part-load operation between full capacity and MSG, the coefficients ηe2 and ηth2 are
different for different CHP sizes, as can be appreciated from Table 7.
A. Economic analyses
The changes in modelling have resulted in changes of operational costs as well as investment
analyses. These changes can be summarized as in Table 6. Results in Table 6 suggest that
changing the formulation resulted in reducing of operational cost, especially the one of DMG
type 5, CHP and TES.
The model of CHP production, defined by (2.2) and (2.3), sets that the CHP efficiencies decrease
when the unit operates in the region bellow maximum rated power. On the other hand, the
objective function "pushes" the CHP to operate as close as possible to the most efficient point in
order to lower the operational cost. Therefore, for DMG type 4 the operational cost will increase
slightly as there is no TES and/or EHP to provide flexibility to CHP operation. Lack of flexibility
means the CHP will operate in less efficient operating point, consuming more gas than in the
previous studies.
50
51
Table 6 Annual operational costs of different dmg types
UNIT TYPE
OPERATIONAL COST (€/a)
Δ TO REFERENCE CASE (%)
TYPE 1
625,440
-
TYPE 2
479,599
-23.3
TYPE 3
445,660
-28.7
TYPE 4
465,130
-25.6
TYPE 5
396,155
-36.6
TYPE 6
339,052
-45.8
TYPE 7
304,217
-51.4
UNIT TYPE
Δ Linear efficiency modelling
(€/a)
TYPE 1
0
Δ Linear efficiency modelling
(%)
0
TYPE 2
0
0
TYPE 3
0
0
TYPE 4
-2,439
-0.54
TYPE 5
53,281
12.61
TYPE 6
3,632
1.06
TYPE 7
7,981
2.56
Results in Table 6 suggest that changing the formulation resulted in reducing of operational cost,
especially the one of DMG type 5, CHP and TES.
For types where additional units are coupled with CHP, these units provide flexibility and enable
CHP not to operate in low efficiency areas thus reducing operational cost. This can particularly
be seen for DMG type 5, where adding TES reduces operational cost by 11% compared to CHP
unit only.
It should be noted that the optimal size of the DMG type 5 changed with the change of the
formulation, optimizing TES size to 250 m3 instead of previous 200 m3. This can be seen in
Table 7 where new optimal sizes are shown. If the same size of storage was kept as in the
previous simulations, the operational savings would be only 5%.
51
52
Table 7 Optimal sizes and relevant efficiencies of units for different DMG DH schemes
TYPE 1
TYPE 2
TYPE 3
TYPE 4
TYPE 5
TYPE 6
TYPE 7
-
-
-
4000
3000
2500
2000
Electrical
efficiency
(full capacity)
0.37
0.37
0.37
0.37
Coefficient ηe1
0.4073
0.4073
0.4073
0.4073
Coefficient ηe2
-315.412
-236.559
-197.133
-157.706
Thermal efficiency
(full capacity)
0.47
0.47
0.47
0.47
Coefficient ηth1
0.515
0.515
0.515
0.515
Coefficient ηth2
-399.891
-310.067
-249.594
-197.417
CHP (kWe)
Boiler (kWt)
6000
6000
6000
6000
6000
6000
6000
Boiler efficiency
0.85
0.85
0.85
0.85
0.85
0.85
0.85
TES (m3)
-
-
200
-
250
-
200
EHP (kWt)
-
5000
3000
-
-
2000
2000
COP
-
3.0
3.0
-
-
3.0
3.0
Optimal size of each DMG is a result of investment analyses and is shown in Figure 21, Figure
22 and Figure 23.
410000
Operational cost (€/a)
405000
400000
395000
390000
385000
380000
375000
150
200
250
300
350
400
TES size (m3)
52
53
380000
370000
NPV (€)
360000
350000
340000
330000
320000
310000
150
200
250
300
350
400
TES SIZE (m3)
Figure 21 Results for sensitivity analyses for DMG type 5
Operational cost (€/a)
380000
370000
360000
350000
340000
330000
320000
2500
3000
310000
3500
300000
1000
1500
4000 CHP size (kWe)
2000
2500
EHP size (kWt)
53
54
1000000,0
NPV (€)
800000,0
600000,0
400000,0
2500
3000
200000,0
3500
,0
1000
-200000,0
1500
4000
2000
CHP siza (kWe)
2500
EHP size (kWt)
Figure 22 Results for sensitivity analyses for DMG type 6
Operational cost (€/a)
350000
340000
330000
320000
310000
300000
290000
300
250
280000
270000
200
260000
1000
1500
150
2000
TES size (m3)
2500
EHP size (kWt)
54
55
1550000
1500000
NPV (€)
1450000
1400000
1350000
1300000
1250000
300
250
1200000
200
1150000
1000
1500
150
2000
TES size (m3)
2500
EHP size (kWt)
Figure 23 Results for sensitivity analyses for DMG type 7
For better understanding, Figure 24, Figure 25 and Figure 26 represent the NPV difference
resulting from different modelling of the electrical and thermal CHP efficiencies. The graphical
values on y axis are expressed as the difference between NPV value calculated with (non)linear
efficiencies and approximations using constant efficiencies (percentage).
NPV modelling difference
90,0%
80,0%
70,0%
60,0%
50,0%
40,0%
30,0%
20,0%
10,0%
0,0%
150
200
250
300
350
400
TES size (m3)
Figure 24 NPV difference in sensitivity analyses for DMG type 5
55
56
NPV modelling difference
80,0%
60,0%
40,0%
4000
3500
3000
2500
CHP size (KWt)
20,0%
0,0%
-20,0%
1000
1500
-40,0%
2000
2500
-60,0%
-80,0%
EHP size (kWt)
Figure 25 NPV difference in sensitivity analyses for DMG type 6
NPV modelling difference
10,0%
9,0%
8,0%
7,0%
6,0%
5,0%
4,0%
3,0%
2,0%
1,0%
0,0%
1000
1500
150
200
2000 EHP size (kWt)
2500
250
300
TES size (m3)
Figure 26 NPV difference in sensitivity analyses for DMG type 7
From the results of NPV sensitivity analyses it can be noticed that linear representation of
efficiency varying with the load results in additional benefits, and therefore NPV in overall, for
all DMG types.
While this change is relatively small for DMG type 6 (around 2%) and around 7% for DMG
57
type 7, the benefits are significantly higher for DMG type 5. This is a result of better utilization
of TES in that particular type. Since the previous approximation of efficiency did not penalize
CHP operation part load, TES was used for storing excess heat only in periods of high electricity
prices. In addition, when efficiency modelling penalizes CHP operation at part load in terms of
lower efficiency and thus higher operating cost, TES is utilized to enable the CHP to operate
longer hours at maximum power. On the other hand, in periods of low electricity prices, CHP can
be switched off and heat demand supplied from TES only. This results in lower operational cost,
as shown in Figure 24, and higher NPV values for the DMG type 5.
However this has little effect on the NPV sensitivity analyses varying the discount rate. DMG
type 6 and 7 remain more favourable choice of investment changing the discount rate from 3, 5, 7
and 10%.
2500000
2000000
NPV value (€)
1500000
1000000
500000
0
3,00%
5,00%
7,00%
10,00%
-500000
TYPE 5
TYPE 6
TYPE 7
Figure 27 NPV dependency on discount rate for different DMG types
Figure 27 compares operational cost difference with regards to efficiency modelling. It can be
seen that for most of the scenarios operational cost differences are below 5%. The exceptions are
scenarios 6 and 9, scenarios where electricity is reduced by 50% compared to the present state.
58
160%
140%
120%
Savings (%)
100%
80%
60%
40%
20%
DMG type 5 savings
DMG type 6 savings
Scenario 9
Scenario 8
Scenario 7
Scenario 6
Scenario 5
Scenario 4
-40%
Scenario 3
-20%
Scenario 2
Scenario 1
0%
DMG type 7 savings
Figure 28 Sensitivity analyses for DMG types 5, 6 and 7
35,00%
30,00%
Difference (%)
25,00%
20,00%
15,00%
10,00%
5,00%
DMG 5 difference
DMG 6 difference
Scenario 9
Scenario 8
Scenario 7
Scenario 6
Scenario 5
Scenario 4
Scenario 3
Scenario 2
-5,00%
Scenario 1
0,00%
DMG 7 difference
Figure 29 Modelling difference sensitivity analyses for DMG types 5, 6 and 7
B. Environmental analyses
With the change of the operational point of the CHP unit, the CO2 emissions from DMG types
4, 5, 6 and 7 will inevitable change as well. From Figure 30 it can be seen that DMG option 4,
CHP unit only, exhibits higher overall CO2 emissions than in the case where constant efficiency
approximation was used for modelling. This can be explained by a lack of flexibility in DMG
59
type 4 operation; CHP unit operates in points with lower efficiency thus using more gas to
produce the same heat and electricity output. This reflects in higher emissions as they are
proportional to the fuel used for CHP operation.
5000000
4500000
CO2 emissions (kg/a)
4000000
3500000
3000000
2500000
2000000
1500000
1000000
500000
0
DMG TYPE 1 DMG TYPE 2 DMG TYPE 3 DMG TYPE 4 DMG TYPE 5 DMG TYPE 6 DMG TYPE 7
2011
2012
CASE A
CASE B
CASE A1
CASE B1
Figure 30 DMG options carbon footprint
30,00%
20,00%
Difference (%)
10,00%
0,00%
DMG TYPE 4
DMG TYPE 5
DMG TYPE 6
DMG TYPE 7
-10,00%
-20,00%
-30,00%
-40,00%
2011
2012
CASE A
CASE B
CASE A1
CASE B1
Figure 31 DMG options carbon footprint modelling differences
In DMG types 5, 6 and 7, where other units provide flexibility in the operation, the emissions
are additionally reduced. This in particular can be noticed for DMG type 5. As explained in the
previous Section, CHP unit is operated only is maximum efficiency points or offline, reducing
60
the operational cost, fuel consumption as well as overall emissions. Adding a TES to the unit has
significant benefits to both minimizing cost and emissions.
With regards to Primary Energy Savings Table 8 shows that changes are insignificant, being
slightly higher for DMG types 6 and 7 (around 2%).
Table 8 Modelling differences in primary energy savings for different DMG options
PPES (UK avg. 2012)
PPES (CCGT)
DMG TYPE 1
0.000
0.000
DMG TYPE 2
0.181
0.309
DMG TYPE 3
0.183
0.318
DMG TYPE 4
0.268
0.065
DMG TYPE 5
0.286
0.079
DMG TYPE 6
0.394
0.289
DMG TYPE 7
0.417
0.317
The final comparison is made for CO and NOx emissions and shown in Table 9.
Table 9 Local emissions assessments for different DMG options
DMG type 1
DMG type 2
DMG type 3
DMG type 4
DMG type 5
DMG type 6
DMG type 7
NOx (t/a)
3.151
0.196
0.117
6.933
7.555
3.394
3.218
CO (t/a)
0.630
0.039
0.023
3.607
4.097
1.887
1.891
The values are slightly higher than in the reference case, where efficiencies are modelled as
constant values, especially for DMG type 6 and DMG type 7. However, the differences are below
2%.
The research and the results shown and elaborated in this and previous Chapter have been
published in [80].
61
FLEXIBILITY - A KEY ELEMENT IN OPERATION AND PLANNING OF
LOW CARBON ENERGY SYSTEMS
Basic operational principle of each energy sector is stable and secure supply of specific energy
vector (electricity, heat, gas etc) procured through various market services delivered at different
time horizons. In order to provide these services, energy systems need to be flexible in order to
maintain supply-demand equivalency, responding to uncertainties and variability in both
production and consumption. The increasing share of renewable energy sources (RES) in the
generation mix redefines requirements on the flexibility of the energy systems. The number of
methods and metrics proposed in the existing literature do not uniquely capture the sources that
can provide flexibility or to what extent do these changes reflect on the operation and planning of
each energy system. The following Chapter presents a comprehensive overview of the operation
and planning flexibility issues, discusses different methodologies proposed for flexibility
quantification. In addition it examines how the increasing need for flexibility in the electricity
sector reflects on interconnected energy sectors, heat and gas, recognizing the value of multi
energy flexibility.
A. Defining flexibility
Goals of reducing global and local greenhouse gas emissions reflect through planning and
operational changes in all energy sectors, in particular electricity, heat and gas. These changes are
mostly due to integration of renewable energy technologies, characterized by their stochastic
nature of production. Although the uncertainty and variability has always been present in these
systems, the integration of RES has increased them, setting new technical and economical
requirements. The share of RES in most countries of the world is still relatively small and the
systems have sufficient flexibility to cope with them; however several systems have already
experienced problems with large share of RES in overall electricity production [81]. These
problems resulted in increased awareness for necessity of advanced planning, tools and strategies
to avoid potential problems. The flexibility in power systems, and similar is valid for all energy
systems, is defined as the capability to maintain constant supply-demand balance due to
uncertainty and variability of those two elements. In addition, flexibility is defined as the
capability of a unit or a cluster of units to respond to system changes over a specific time horizon
62
as each unit is constrained by its technical capabilities. The definition of flexibility should also
include where and how electricity is produced, how it is transported, how and when it is traded
and when it is consumed [82].
System flexibility is of high value to the system operators as they have a task of maintaining
system stability; on the other hand it has little to no importance to power suppliers whose main
interest is making profit. This creates a challenge when it comes to incentives for operational
flexibility providers (e.g. for units to ramp up or down due to variability in RES production) and
defining benefits of investing in new flexible technologies. There are significant efforts to
research the capability of the existing power systems to deal with increasing share of wind and
solar generation and to define and evaluate the flexibility to smoothen the uncertainty and
variability these sources bring to the usual operation practices. Figure 32 visualizes how the need
for flexibility increases with integration of RES in the system. The arrows in Figure 1 point to the
most obvious timeframes when there is a potential lack of flexibility in the system: large changes
in net load often referred to as ramping, and excess wind production requiring operational
changes of slow, inflexible generation such as nuclear. During those periods it is often infeasible
to shut down inflexible generation and the operators rather chose curtailing wind energy (in cases
when exporting excess electricity to neighbouring balancing area is not possible).
To address the challenge of significant wind integration new concepts for electricity generation
dispatch through different unit commitment (UC) problems modelling are being developed,
recognizing benefits of stochastic and rolling UC approaches, emphasising the value of load and
RES forecasting [83]. A group of authors from National Renewable Energy Laboratory presented
a detailed analysis of connecting balancing areas (BA) in the US, researching ramping
capabilities in time horizons specific for those BA [84]. The authors elaborate on a unified
standard set by North American Electric Reliability Corporation (NERC), defining Area Control
Error (ACE) as frequency deviation from set value and introducing minute values and monthly
average for statistical evaluation of ACE; called Control Performance Standard (CPS1 is defined
on a minute resolution and CPS2 as monthly average).
63
2500
Electricity (kW)
2000
1500
1000
500
16:00
0:00
8:00
16:00
0:00
8:00
16:00
0:00
8:00
0:00
16:00
0
Time (h)
Nuclear
Coal
Gas
Hydro
Load
Increased flexibility needs
3000
Electricity (kW)
2500
2000
1500
1000
500
16:00
0:00
8:00
16:00
0:00
8:00
16:00
0:00
8:00
16:00
0:00
0
Time (h)
Nuclear
Coal
Gas
Hydro
Wind
NetLoad
Figure 32 Flexibility requirements change with increasing share of RES
In [85] the same authors discuss how the ramping requirements and wind integration cost are
reduced if BAs cooperate in balancing electricity deviations. The report elaborates that, when
comparing ramping requirements before and after wind integration, the benefits of coordinating
several BAs are more emphasised after wind integration. With no wind, morning and evening
requirements for ramping up and/or down coincide and these benefits are less pronounced. The
analyses were conducted on a 5-minute scale and to demonstrate these benefits the report
introduced several metrics for each system.
a) dominant ramp; defined as the maximum up and down ramp requirement at a given hour,
64
b) secondary ramp; every ramp requirement of opposite direction to dominant ramp at a
given hour,
c) ramp penalty; the difference between the dominant ramps required by separate systems
operation and combined operation.
Ramping requirements are defined separately for each of the 4 BA-s separately and resulting in
the definition of the metrics for BAs separate operation. Ramping penalty is used as a metric
demonstrating how these requirements reduce when BAs operate in a coordinate fashion. It
should be noticed these metrics are hourly average values and do not accurately present the need
for flexibility, however they provide a good insight into benefits on interconnections and joined
BA operation. The authors use a rather simple approach; they extract the ramping capability of
each thermal unit based on the available past operation of the units. While the results suggest that
presently ramp down capability is more significant than the ramp up as most units operate close
to the maximal point, this will change with higher penetration of RES as the units will not operate
so close to the maximum due to increased reserve requirements. In addition they conclude that
balancing areas where sub-hourly markets exist pay little or nothing for ramping capability of the
generators (the little or nothing refers to average cost of energy). If the generation units have
sufficient physical capability to ramp up or down and there is an adequate sub-hourly energy
market structure the generators are incentivized to respond to net load changes. System operators,
particularly in the US, report on the annual insufficient capability to provide ancillary services as
scarcity events [86] and put particular emphasis on insufficient ramping capabilities of the system
to respond to an unpredicted event. A methodology for reducing these events and ensuring that
additional ramping capability is available is elaborated in [87], proposing a more robust market
dispatch model. The proposed concept results in slightly higher prices at the real-time market (in
the presented case real-time market is settled 5 minutes before event); however is also eliminates
scarcity events. Overall, the authors of the paper claim their approach will in long term result in
lower cost as it will eliminate ramping price spikes.
In the presence of high wind production, utilities face a problem of system minimum generation
requirements, mostly expected during overnight hours. As mentioned previously and shown in
Figure 32, in those the operators usually decide to curtail wind production. This has already been
observed in Denmark as a consequence of its reliance on combined heat and power electricity
plants for district heating and minimum generation constraints of those units [88]. The issue of
65
minimum generation constraints will require different approaches to operation practices in the
future high RES power systems, including market changes or modifying technical characteristics
of the generator in order to cycle bellow current minimum stable generation points [89]. A single
change will probably not suffice as several markets have already observed cases where energy
prices have dropped below the fuel cost of producing electricity, resulting in negative energy
prices and high positive regulation prices [90]. This has shown that power plants choose to lose
money paying to sell energy rather than to further reduce their output and shut down as such
actions would reflect in very high losses (in case where they would have to reduce their output
bellow Minimum Stable Generation - MSG). This suggests that flexibility of the systems will
also depend on the power plants ability to quickly shut down and start up again.
B. Sources of flexibility
Available literature categorizes sources of flexibility differently, but the most common is the one
found in [91] which defines generation flexibility, imported flexibility and demand side
flexibility. Similar to the mentioned report the following section elaborates on research areas for
each of the defined flexibility sources and capacities. Figure 33 conceptually presents these
flexibility sources.
66
LOW FLEXIBLE AND
INFLEXIBLE
GENERATION
•
•
NUCLEAR
COAL
INFLEXIBLE
CONSUMPTION
FLEXIBLE GENERATION
UNITS
•
•
•
•
GAS
HYDRO
CCGT
CHP
MARKET
SERVICES
FLEXIBLE
CONSUMPTION
IMPORTED FLEXIBILITY
•
•
•
•
•
EV
STORAGE
DR
AC LINES
HVDC
Figure 33 Flexible and inflexible sources
B.1.
Storage
In the context of storage, latest research focuses on "storing" electricity in battery storages. For
this reason the paper makes a distinction between battery storages and thermal or hydro pumped
storage and elaborates only on battery storage in this section and refers them as storage.
Different technologies are suggested to have the potential for providing uncertainty and
variability balancing in high RES power systems. However, services these units can provide are
limited due to the technical constraints: energy (time for service provision) and power [92]. In
[93] the role and potential services energy storage could provide are elaborated emphasizing the
economy behind those services. In the context of the report, energy storage is observed as an
alternative to wind curtailment since energy storage has the capability to reduce the overall
system minimum generation (MSG) constraints. In [93] a metric to evaluate the perspective of
the energy storage is introduced, called Flexibility Factor (FF), defined as a ratio of annual peak
and the systems minimum generation point. For most systems system level MSG is usually
around 30-40% which corresponds to FF of 60-70%. As the share of RES increases minimum
generation point will have to be lower in order not to curtail wind; these benefits create, among
67
others, a business case for energy storage. In addition, with respect to percentage of RES
integration, the authors distinguish two types of flexibility: a) ramping flexibility: needed at
lower levels of RES (following net load variations); b) energy flexibility: needed for higher RES
share, and expressing it as the ability to increase the coincidence of RES generation and
electricity demand. A more detailed elaboration of this can be found in [94]. In case of Europe,
several studies give very similar answers to the role of energy storage in the high RES power
system, however this is highly dependent on the systems energy mix of a specific country [88],
[95], [96]. A general conclusion shown also in [97] is that energy storage can alleviate the
problems of the uncertainty and variability of renewable energy sources. With the present cost of
energy storage, these units will have to provide several services to increase the overall benefits
justifying their investment [98]. For some of those services, such as frequency control, temporal
arbitrage, provision of reserve, the location of the storage does not significantly affect the value
that it provides. On the other hand, transmission or distribution congestion services highly
depend on the position as well as sizing of the units and respectively feasibility of a specific
storage unit. Since they are distributed, these storage devices can perform a spatiotemporal
arbitrage preventing network congestion and curtailment of RES which results in reducing the
cost of producing energy from conventional generating units. In [99] the authors extensively
elaborate on storage capability to provide a specific service. They define 9 characteristics
constraining the storage capability to provide a specific service, additionally dividing them into 6
physical characteristics and 3 scenario dependant characteristics. The paper argues storage can
provide a number of services to different power system stakeholders, dividing those services
according to the speed and frequency those services are required as well as by the system subject
potentially requiring a specific service. It is interesting to notice that while a number of papers
elaborate on storage capability to provide one or more services to a single stakeholder [100],
[101], [102], [103], there is a limited number of those recognizing that additional value of storage
lies in providing multiple services to multiple market/system subjects. An unique approach is
presented in [104] where storage units offer services to multiple system players over different
scheduling periods. Storage is modelled as a provider of flexibility through various market
services procured in consecutive tenders, setting storage energy equilibrium constraints over a
period of time for each service offered. This ensures the process can be repeated for each service
during the following settlement period. The longest duration settlement period is cleared first and
68
the decisions are passed on to the next settlement period as constraints. It should be mentioned
the entire analysis is based for a very large storage (size being 25% of the peak power of the
analyzed system)
B.2.
Flexible generation
Conventional generation units designed to have more flexible characteristics like higher ramp
rate (MW/min), larger operating range (Pmax/Pmin), smaller start-up/shut-down times and
minimum up and down times are usually considered as the low cost sources of providing
flexibility. A very useful overview of the technical characteristics of existing conventional
generation units in context of their ability to provide flexibility is given in [105]. In [106]
flexibility of single units, and correspondingly the entire system, is defined by 4 parameters: i)
Magnitude - defined as the size and the direction of change, taking into account MSG; ii) Ramp
response - defined as the rate of change; iii) Frequency - defined as the number of events; iv)
Available Flexible Resources - defined as the ability to change between resources. In the
operating timeframe, two defined metrics for flexible resources are operating reserve (usually
divided into two or more categories based on speed of response) and system regulation
(automatic generation control). With respect to that, the report creates mapping of various events,
going from large magnitude events with a high ramp requirement to high frequency and ramp
response but with low magnitude [107].
The proposal in [108] suggests that a concept based on multiple smaller units with coordinated
operation, at a single location, have the capacity to provide fast response, capacity and energy to
the operator. Several examples already exist worldwide. The idea is similar to a known concept
as Virtual Power Plant [109].
B.3.
Demand response (DR)
Renewable energy integration and increasing variability and uncertainty in the power system
imply that fewer subjects in the system will stay passive; in particular this relates to the
consumers. Traditionally, consumers do not change their behaviour or consumption with respect
to system or market signals, although their active participation is an attractive topic for over two
decades [110]. Responsive consumers are recognized as a potentially valuable source in
providing various system services, going from frequency response [111], reducing wind forecast
69
errors [112] or joint energy and reserve services [113]. Such active participation responding to
system signals creates another valuable source of flexibility in the future power system. In [114]
the authors discuss potential barriers for active participation of aggregated flexible demand in
Nordic markets. The authors make a very relevant statement in terms of flexibility; spot market
prices cannot be price signals regulating flexible consumption, as often suggested in the
literature, due to the limitations in metering and communication infrastructure. However,
aggregated consumers can offer balancing regulation based on intra-day prices. Consumers
aggregated in such way are seen by Transmission System Operator (TSO) as a single market
Balancing Responsible Party (BRP) and often referred in the literature as a Virtual Power Plant
(VPP) [115]. Aggregating consumers in a single stakeholder, VPP, creates a market entity
capable of quickly reacting to system demands. However, such entity has high power capacity to
provide services in the ancillary services market, such as primary and secondary reserve, but
significantly lower energy capacity. Authors in [116] address this issues, focusing on the
capability to provide primary and secondary reserve service. The flexibility area is defined as a
difference between baseline consumption and the total flexible demand consumption taking into
account forecasting error for the baseline prediction. The authors propose a simple formulation:
∂E (t )
= P cons (t ) − P base (t )
∂t
(6.1)
Where energy derivation presents the energy available from flexible demand, Pbase is the baseline
consumption presenting the inflexible consumption and each deviation expressed as Pcons is the
provision of ancillary services from the VPP. The consumption profile is defined by its maximum
and minimum from where the capacity for service provision can be extracted:
P cap =
min( P max − max( P base (t )), min( P base (t )) − P min )
t∈T
t∈T
(6.2)
In [117] responsive consumption is divided into sheddable, controllable and acceptable load as
means to quantify the capability of a specific consumer to participate in DR services and to
define how much flexibility can a specific consumer provide. Using this methodology the authors
70
define the total flexibility metric for DR as a product of sheddable load and minimum between
acceptable and controllable load, as these two are in fact coincident. These three metrics
are expressed as percentages of load profile in each observed time step. The results show that
DR has a significant capability (participates with over 90%) to provide reserve regulation for
studied cases. In addition to the literature referenced above, in [118] it is recognized that in cases
where there is a need for a larger response, in terms of power relative to energy of the load, it
might impact the functionality these devices offer to the end users (for example the state-ofcharge of the EV battery). Discussing different strategies for control and different aspects of
providing demand flexibility, the authors conclude hierarchical model of control is the most
promising one. Again the conclusions are similar to the conclusions made on the aggregating and
control of distributed resources in the context of VPP; the concept of such control is
demonstrated in Figure 34.
B.4.
Electric vehicles (EV)
Today’s transport system is responsible for 13.5% of greenhouse emissions worldwide, and
personal road vehicles account for 44.5% of that share [119]; it is easy to conclude that only
personal vehicles are responsible for at least 6% of greenhouse gas emissions worldwide. One of
the solutions is substituting the internal combustion engine with an electrical one. This makes
sense to a certain extent as the electricity sector is strategically oriented towards renewable
energy sources, thus the negative ecological impact of electric vehicles (EV) that would use
electricity produced by renewable sources, would be insignificant. Having a fleet of flexibly
chargeable EV can be regarded as aggregated flexible demand and as such the benefits of EV
flexibility are often discussed in the context with other demand responsive units such as electric
heat pumps [120]; both considered to have a high potential to create new load peaks at the
domestic level.
71
SYSTEM CONTROL
VPP 1
DG
aggregator
Load
aggregator n
Load
aggregator 1
VPP 2
DG
aggregator
Load
aggregator 2
VPP n
DG
aggregator
Load group
Load group
DG group
DG group
DG group
Load group
POWER SYSTEM
Figure 34 Hierarchical control of DR and VPP
A major difference from what is considered classical flexible demand, usually thermal load
devices, is that EV have much higher flexibility in terms of temporal scheduling as their charging
can be postponed for several hours without significant impact on the service to the user. In
addition, under the concept of Vehicle-to-Grid (V2G) EV would be capable of injecting
electricity back into the system and providing a number of ancillary services such as primary
frequency response [121], spinning reserve [122], load and voltage management [123] and even
be able to provide both energy and ancillary services maintaining preferable operation point
(POP) within satisfactory boundaries [124]. As flexible loads, EV are scheduled at the energy
market to increase the net load (demand reduced by the produced RES electricity) of the system,
often referred as valley-filling, with the goal of minimizing part load operation or expensive shut
down/start up of the generators. Following on this, the authors in [125] research the impact EV
will have on future energy generation mix. New unit commitment algorithm considers retirement
of specific power plants in predetermined schedule and analyzes multiple scenarios to define the
optimal investment into future generation mix. Although the conclusion made is that higher EV
integration will stimulate investments into more flexible, but also for the system more expensive,
power plants the authors also argue the conclusion cannot be generalized. Consumption
seasonality and CO2 emission pricing will have a high impact on EV pricing and charging and
thus future generation portfolio investments.
72
Dual energy and ancillary service approach is presented in [126] where EV are presented
twofold: as V2G and as flexible demand only. In both cases they can provide ancillary services as
gradual or as interruptible charge/discharge units. This research seems to be the only paper
focusing on the possibility of EV to provide tertiary reserve, or flexibility in terms of energy, to
the system. Single aggregator is considered as a coordinator of EV fleet, however the probable
concept for control and flexibility provision of EV would resemble to the concept shown in
Figure 34.
C. Quantifying flexibility
In addition to modelling system operation, several papers have recognized value in defining an
"offline" metric in order to define the capability of the system to adapt to changes and evaluate
the flexibility of the power system. Very roughly these can be categorized as operational and
planning metrics [127], however the following Section will elaborate further on time frame
categorizations, specifically in terms of operational flexibility.
C.1.
Operational flexibility
In case of a disturbance in the system, resulting in deviation of the frequency from its nominal
value, the generating units immediately respond in order to maintain the balance - this is called
primary frequency response. Traditionally this is a feature of conventional generators. Although
this response is automatically provided by the units, some recent research showed a decline in the
frequency response capability of the units. There are several reasons for this, starting from poor
incentives as the market for this service practically does not exist to reserving the capacity for
other, more market subsidized services such as secondary or tertiary reserve. In addition,
governor ramp constraints are often not included in the unit commitment dispatch neglecting
units' capability to adequately respond to primary frequency. While in [128] the authors
demonstrate s simple way to include this into generation dispatch, a comprehensive approach to
system unit commitment modelling including primary, secondary and tertiary response service is
in detailed elaborated in [129], [130]. The authors recognize that such a systematic evaluation of
the flexibility for larger systems is computationally too extensive and requires a large amount of
often unknown data on all units in the system.
73
How specific market operates and in what time frame the generators are dispatched has a high
impact on regulation and the flexibility of the system [131]. In California several scheduling
time frames (day ahead (DA), hour ahead, real-time market) with load and variable generation
forecasting data, define the needed regulation as a difference between scheduled and actual
generation/load over those time frames. The authors define the needed flexibility as a triad of i)
power capacity (MW), ii) ramping capability (MW/min) and iii) energy (MWh) [132]. Taking
into account forecasting error, a method of "flying brick" is used to evaluate the available
balancing reserve of the generators on a DA and hour-ahead basis, defining periods when
additional generators need to be committed or decommitted. In [133] the same authors propose a
methodology for recognizing harmers and helpers to system imbalances and a methodology to
assign the cost of balancing to each unit. The methodology is based on metrics defined by NERC,
namely Control performance standard (CPS) and balancing area control error (ACE) [84]. The
methodology only considers energy, not the technical capabilities of each unit, thus not providing
a metric for evaluation of flexibility but rather a post event evaluation of units that contribute to
balancing the system or cause deviation from scheduled operation. CPS performance is analyzed
in [134], evaluating impact of large wind integration on the performance standards and proposing
solutions in terms of increased forecast accuracy, combining control areas and increasing
frequency response reserves. Furthermore, in [135] needed improvements are evaluated based on
the estimated variability each new non-controllable unit, meaning wind in the mentioned work,
contributed to the existing generation fleet.
Effective Load Carrying Capability (ELCC) index is defined in [136] as a measure of individual
new generators contribution to the overall system level of reliability, calculating the later one as
the reliability for the system in the previous year. This metric serves as an indicator, value, of the
load that can be served at the same level of reliability as before integrating variable RES into the
system. In essence it tells if the net load can still be served after adding a new RES unit. This
metric is recognized by the Working Group on Variable Generation Capacity Value as a very
promising one [107].
Loss of Load Expectation (LOLE) is a known methodology for system reliability evaluation
which authors in [137] use to create a flexibility metric called Insufficient Ramping Resources
Expectation (IREE). This metric is later expanded [138] and shown in (6.3) and (6.4):
74
=
FLEX t ,i ,r Online Onlinet ,r * min( RRr * i, RatedCapacityr − Pr oductiont ,r )
FLEX t ,i ,r Offline =
min( RRr *(i − StartTimer ), RatedCapacityr *(1 − Onlinet ,r ))
(6.3)
(6.4)
From the flexibility metrics expressed as in (6.3) and (6.4) the IREE is calculated with the goal of
evaluating flexibility of each unit based on its ramping capability (RRr) and minimum stable
generation constraints, taking into account if the unit is online, over a period of time i. The IRRE
value is the expected number of times a system will not be able to meet net load change during a
defined time horizon. The metric also eliminates all the cases where the available flexibility is on
the upper required limit by reducing it by 1MW. The conclusion made is that positive IRRE
maximum values coincide with the start up times of flexible resources while negative are driven
by net load change magnitude. Since the metric requires operational points of each unit, the same
group of authors suggests a simplified approach, where these operation points are based on merit
order [138]. The concept is further simplified by introducing flexibility metric of online available
units and flexibility metric detecting offline units which can be started in a needed time frame.
These values can then be used to calculate the flexibility deficit. An interesting approach
borrowed from the control theory is presented in [139]. The flexibility index is expressed as ratio
between cut and uncut polyhedron where the uncut polyhedron presents the entire operating
region while the cut polyhedron presents the feasible area of operation defined by the technical
constraints. The methodology is rather complex and not elaborated in enough details to determine
the potential for practical application.
C.2.
Planning flexibility
Although it is difficult to draw a strict line and decouple operational and planning flexibility (as
they are tightly related), the following subsection elaborates on models and metrics examining
the flexibility, and sources to provide this flexibility, in future scenarios usually targeting a
specific energy mix goal. In [140] the authors propose an offline metric to evaluate the flexibility
of each unit in order to optimally plan a future flexible system defining the entire system
flexibility as a sum of each unit's flexibility share. This metric is shown by (6.5) and (6.6):
75
1
1
[ Pmax (i) − Pmin (i)] + [ Ramp(i) * ∆t ]
2
flex(i ) = 2
Pmax (i )


Pmax (i )

FLEX A = ∑
* flex(i ) 


i∈ A ∑ Pmax (i )
 i∈A

(6.5)
(6.6)
This metric enables evaluation of unit's flexibility without the need for knowing the operational
point of it; it is based on the range between maximum power (Pmax) and MSG (Pmin) and ramp
rate of the unit (Ramp(i)). To evaluate the investments into new flexible generation unit, a new
Unit Construction and Commitment (UCC) algorithm is developed as an extension of the one
presented in [141]. Decisions on the new generation investment are based on profitability, and
therefore flexibility, of unit's participation in the day ahead and real-time balancing market,
enabling generators to sell both energy and reserve (receive an option fee for reserve but, if
deployed, also exercise price based on real-time balancing market). The UCC performs a rolling
scheduling to update the generation dispatch every 6 hours with new load and wind forecasts for
three different sets of generation mix. The paper also introduces normalized profit metric in order
to determine the profit each generation unit can gain by deploying flexibility services (in realtime balancing market). Similar model is presented in [142], however with a lot less details in
terms of modelling constraints or quantifying flexibility through a specific metric. The paper
evaluates feasibility of installing Compressed Air Energy Storage (CAES) systems in the area of
Bretagne in France in order to balance wind fluctuations for two scenarios in year 2030.
Although this research can be considered as case based due to specifics of the French power
system (inflexible nuclear with high minimum stable generation and significant share of variable
and uncertain wind generation) the paper is interesting from the perspective that placement of
CAES is considered taking into account geographical features as well as economy and system
needs. Concept proposed in [143] takes into account adaptive cost as well as the operational
flexibility by an adjusted unit commitment model. The authors argue that the solution for future
generation mix, subject to high levels of uncertainties, has to be able to adjust to all possible
future scenarios and that the obtained solution has to have minimum cost if adjustments are
76
needed due to imperfect scenario knowledge. With this in mind, the multi-stage model proposed
defines a flexible system taking into account global and local uncertainties.
An relocation coefficient Rc is introduced in [81], further elaborated in [144], and defined as a
statistical correlation between net electricity exchange between unit and the system and the net
electricity demand of the system. The term is introduced to demonstrate how a combination of
CHP and heat pump (EHP) (or any selected) can be more beneficial to the system with high share
of wind compared to CHP only units. It is not entirely clear how the values for assessing Rc are
obtained but they present an average annual value. The metric is shown in equation (6.7):
Rc =
∑ (e − e )(d − d )
∑ (e − e ) ∑ ( d − d
m
2
m
m
2
m)
(6.7)
In the equation above, e values present the exchange between a generation unit and the system,
while d is the net load demand. The same factor is mentioned in [36], called marginal "goodness"
of a plant and further defines that a total system has a value of 1.0 while a RES unit would have a
negative value. The concept is called intermittency friendliness and compares different options
with respect to mentioned factor. It should be noted that the metric is a statistical annual average
and does not reflect flexibility of the system, as it does not capture intra-hour changes in the
system. However, the authors suggest that it shows how flexible the unit is in responding to
system requirements. If the unit has an Rc of 1.0 it can operate in a standalone regime.
D. Multi-energy system flexibility aspect
Despite the above mentioned work, there is still a lack of a comprehensive methodology to
evaluate the flexibility of the future high RES power system and how increasing this flexibility
will reflect on other, interconnected, energy systems such as gas or heat. Interactions between
energy vectors occur on multiple levels and in the system. An example of such coupling are
Combined Heat and Power (CHP) plants which can be regarded as energy hubs coupling three
vectors; gas, heat and electricity [145]. Gas fuelled power plants are commonly used to balance
the variations in RES production meaning that increased share of RES in the system will reflect
the gas system economics and imbalance costs [146], [147]. However CHP units rarely decouple
their operation from demand of one vector. For example, CHP will usually follow heat demand,
77
using gas when needed and producing electricity as a by-product. Adding thermal energy storage
(TES), in cascade with CHP, means that heat production will not anymore be driven only by
demand. In that case TES has a role of buffer and CHP can react according to market signals and
act, at a household level, as an demand response unit [65]. Adding TES to CHP creates a flexible
unit, capable of reacting to power system demands and delivering the needed heat demand to the
final consumer. In [148] the authors try to quantify the amount of flexibility that can be gained
from coupling CHP and TES. The concept is based on the capability to push the CHP to operate
at its maximum when there is a need for upward reserve in the system, and to keep the CHP
offline when there is a need for downward reserve in the system. However the metric used to
quantify this flexibility is expressed as time the unit can be pushed to operate in the extreme
conditions, partially ignoring technical constraints of the CHP unit such as ramp rate, MSG of
minimum up/down time.
The planning concept of all energy vectors simultaneously has gained a lot of attention of the past
year, see for example [46], [6], however the interactions in terms of operation and flexibility have
not been addressed properly. The concept of Multi-Energy systems [5] suggests there is high
value in creating an integrated energy system where interaction of different energy vectors can
bring additional value in terms of economic and environmental benefits [149]. General concept of
Multi-energy systems interactions are presented in Figure 35.
78
ITL
BSS
EHP
∑
TS
IEL
∑
SOLAR
WIND
FTL
CHP
AB
FEL
PHEV/EV
BSS
à
EHP
à
CHP
à
FTL
à
ITT
à
AB
à
FEL
à
IEL
à
PHEV/EVà
TS
∑
Battery Storage System
Electric Heat Pump
Cogeneration unit
Flexible thermal load
Inflexible thermal load
Boiler
Flexible electric load
Inflexible electric load
„Plug-in” hybrid electric
vehicle/Electric vehicle
Thermal storage
à
Heat
Gas/Fossil fuel
Electricity
Energy flow
Figure 35 Flexibility from energy vector interaction
The sum boxes present the controllers which would, interacting with different units, be able to
provide flexibility from multiple energy vectors. In [32] the authors present the concept of energy
shifting, elaborating that each energy vector can be provided from multiple sources in flexible
multi-generation unit, for example heat can be delivered from CHP, boiler or electric heat pump.
By intelligently responding to system demand the unit can shift the operation of each component
in the multi-generation unit and produce the desired energy vector. The concept is further
developed and has shown to be effective as a flexible demand response on small scale level [150]
and as a provider of ancillary services [35]. A general concept where different energy vectors
interact in order to provide flexibility is shown in Figure 36.
79
Current Energy Systems
Hydro Power
plant
Demand
Thermal
Power plant
Nuclear
Power plant
Electricity
Combined Heat
and Power
Renawable
Energy Sources
Heat
Boilers
Fossil Fuels
Transport
system
Future Flexible Energy Systems
Flexible
demand
Hydro Power
plant
Thermal Power
plant
Nuclear Power
plant
Flexible T&D system
Electricity
Battery
Storage
Thermal
Storage
Combined
Heat and
Power
Electric
heating
Renawable
Energy
Sources
Heat
Boilers
Fossil Fuels
Plug in Hybrid
Electric
Vehicles
Transport
system
Figure 36 Current and future flexible energy systems
80
E. Operational flexibility metric of Distributed Multi-Generation
The only known concepts capturing interactions between multiple energy vectors are described in
in [35] and [150]. Through a defined concept of shifting factors, these papers proposed that
flexibility can be gained from shifting to another energy vector. To be more precise; the
alternative for delivering the desired output to the end consumer can be achieved by choosing an
alternative component, from DMG unit to produce this. The newly selected component uses
different input energy vector but produces the same output vector. The concept will be elaborated
further by Figure 37.
Eimp
Eexp
(1-α)Echp
EDS
Fchp
ED
Ehp
αEchp
CHP
ηe, ηt
Electric Heat
Pump
COP
Hchp
TES
Hs
Haux
Faux
Hhp
Auxiliary boiler
ηaux
a)
C
O
N
S
U
M
E
R
S
ED, HD
81
Eimp
Eexp
(1-α)Echp
EDS
Fchp
ED
Ehp
αEchp
CHP
ηe, ηt
Electric Heat
Pump
COP
C
O
N
S
U
M
E
R
S
ED, HD
Hhp
Hchp
TES
Hs
Haux
Faux
Auxiliary boiler
ηaux
b)
Eimp
Eexp
(1-α)Echp
EDS
Fchp
ED
Ehp
αEchp
CHP
ηe, ηt
Electric Heat
Pump
COP
Hchp
TES
Hs
Haux
Faux
Hhp
Auxiliary boiler
ηaux
C
O
N
S
U
M
E
R
S
ED, HD
c)
Figure 37 Flexibility provision a) optimal operation points, b) upward flexibility requirement, c) downward
flexibility requirement
In Figure 37 a) the optimal operation point flow diagram of DMG type 7 is shown. A similar can
be presented for each type; however type 7 is chosen as it is the most comprehensive one. In
previous chapters it was elaborated that the optimal operation points are calculated day ahead,
defining the behaviour of the each unit over a 24 hour horizon resulting in the minimal operation
cost. In cases when there is a need for additional flexibility, DMG units can provide this in a
82
unique way without affecting neither the contracted electricity exchange from the day ahead
market or final consumer demand. The second issue is related to the idea of demand response;
however it should be noticed that, unlike classical concept for DR provision, DMG units can
respond to system requirements without affecting the quality or quantity of service for the
customer, in any way.
In Figure 37 b) the concept for upward reserve flexibility is demonstrated. Darker colours, in
comparison to Figure a), are depicting additional production of the components final energy
vector, while lighter colours show the opposite. The upward flexibility for DMG type 7 is
provided by:
•
CHP unit production additional electricity. As a consequence, more electricity is exported
to the system, but also more heat is produced which needs to be stored in TES. This
means TES is, alongside CHP maximal capacity, additional limiting factor is such
flexibility provision. In case TES is the limiting factor, the heat production of boiler (if
possible) should be reduced.
•
EHP unit reduces its operation. The EHP acts as electrical demand and by reducing its
production it directly impacts the amount of electricity exchanged with the system. The
limiting factors in this case are the difference from the operational point of EHP and
shutting down the unit, and the amount of heat not produced by EHP that can be replaced
by producing it in the boiler discharging it from TES.
The above described behaviour or regulating the operation of EHP can be described with the
following equations (6.8) and (6.9):
=
H EHP DOWN (t ) min( H EHP (t ), H AUX _ MAX − H AUX (t ) + H s (t ) − H S _ MIN )
EEHP DOWN (t ) =
H EHP DOWN (t )
COP(t )
(6.8)
(6.9)
The H EHP DOWN (t ) is the maximum amount of heat that will be shifted by reducing the operation
of EHP. This heat needs to be replaced by producing it from another source. The available
sources for production are TES and boiler. The basic idea would be to substitute the heat from
83
EHP by discharging TES. Only if this is not sufficient, boiler should be additionally started to
cover the difference.
In case when EHP is the limiting factor, minimum of the two values in equation (6.8), TES will
discharge the amount of heat equal to HEHP, to substitute for EHP production. In case this is not
sufficient, the boiler will substitute for the rest. The new set point of TES needs to be calculated
as in (6.10):
H
=
max( H S (t ) − H EHP (t ), H S _ MIN )
S _ NEW (t )
(6.10)
Defining the storage set point is relevant as this is the limiting factor in pushing the upward
operation of CHP unit. This is then described by (6.11) and (6.12):
H CHPUP (t ) =min( H CHP _ MAX − H CHP (t ), H S _ MAX − H S _ NEW (t ))
ECHPUP (t ) = H CHPUP (t ) *
ηe (t )
ηt (t )
(6.11)
(6.12)
H CHPUP (t ) is the maximal additional heat that will be produced by pushing upward operation of
the CHP unit, as heat is produced together with additional electricity needed by the system. Total
available upward flexibility is defined by equation (6.13):
UP
=
FLEX
(t ) ECHPUP (t ) + EEHP DOWN (t )
(6.13)
In Figure 37 c) the concept for downward reserve flexibility is demonstrated. Darker colours, in
comparison to Figure a), are depicting additional production of the components final energy
vector, while lighter colours show the opposite. The downward flexibility for DMG type 7 is
provided by:
•
CHP unit reduces its electricity output. This results in less electricity exported to the
system and less heat produced. Limiting factor is providing downward flexibility from
CHP is the MSG of CHP unit. The amount of heat not produced needs to be recovered
84
from storage; in cases this is not possible, the operation of the boiler will be increased in
order to substitute for the missing heat.
•
EHP unit increases its operation and thus increases electrical demand; this will impact the
amount of electricity exchanged with the system, increasing the import from the system.
The limiting factor in this case is the difference from the operational point of EHP and
maximum installed thermal power of the EHP unit. The amount of additional heat from
EHP needs to be stored in TES. If this capacity of TES is lower than the difference
between EHP operating point and installed EHP capacity, production from boiler will be
decreased.
The above described behaviour or regulating the operation of EHP can be described with
equations (6.14) and (6.15):
UP
H EHP=
(t ) min( H EHP _ MAX − H EHP (t ), H AUX (t ) + H S _ MAX − H s (t ))
EEHPUP (t ) =
H EHPUP (t )
COP(t )
(6.14)
(6.15)
In case when EHP is the limiting factor, minimum of the two values in equation (6.14), TES will
be additionally charged with the amount of heat equal to HEHP. In case the capacity of storage is
not sufficient, the boiler will reduce its output for the required difference. The new set point of
TES needs to be calculated as in (6.16):
H=
min( H S (t ) + H EHP (t ), H S _ MAX )
S _ NEW (t )
(6.16)
Additional downward flexibility is achieved by reducing the output from CHP. Limiting factors
in this case are CHP MSG level and the amount of heat in TES. Additionally heat can be replaced
from boiler. This can be described by equations (6.17) and (6.18):
DOWN
H CHP
=
(t ) min( I CHP (t )*( H CHP (t ) − H CHP _ MIN ), H S _ NEW (t ) + H AUX _ MAX − H AUX (t ))
(6.17)
85
ECHP DOWN (t ) = H CHP DOWN (t ) *
ηe (t )
ηt (t )
(6.18)
The H CHP DOWN (t ) is the maximum amount of heat that will be shifted by reducing the operation
of CHP unit. Total downward flexibility available by DMG unit is described by (6.19):
DOWN
FLEX
=
(t ) ECHP DOWN (t ) + EEHPUP (t )
(6.19)
The above elaborated concept of flexibility is presented in Figure 38-Figure 40 for DMG type 7
for specific days of the year, winter, spring/autumn and summer. The flexibility values for each
day are calculated based on optimal values obtained in simulations in Chapter 4.
Flexibility metrics can be, among others, used to evaluated potential additional services DMG
units can provide. While majority of system services such as ancillary services highly depend on
the time when their utilization is negotiated, an interesting aspect to observe is the capability of
DMG units to independently operate from the grid. For this reason figures compare the calculated
flexibility from shifting with imported electricity (if it is upward flexibility) and exported
electricity (if it is downward flexibility).
As it can be seen from Figure 38, during all periods of a winter day DMG unit has sufficient
flexibility to substitute for the electricity exchanged with the grid. In theory this means that at any
moment of the winter day, the DMG unit can work as an independent microgrid system, being
completely separated from the system. Additional dynamic simulations could be conducted to
find if the system can provide a valuable system service such as sheddable load in demand
response programs, assisting the power system in critical moments.
86
DMG TYPE 7 - FLEXUP
Electricity (kW)
3000
2500
2000
1500
1000
500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
FLEXUP (kW)
Imported electricity
DMG TYPE 7 - FLEXDOWN
Electricity (kW)
2500
2000
1500
1000
500
23:00
22:00
21:00
20:00
19:00
18:00
17:00
16:00
15:00
14:00
13:00
12:00
11:00
10:00
9:00
8:00
7:00
6:00
5:00
4:00
3:00
2:00
1:00
0:00
0
Time (h)
FLEXDOWN (kW)
Exported electricity
Figure 38 DMG type 7 additional operational flexibility for a winter day a) upward, b) downward
Similar findings as for the winter day can be made for spring/autumn day as shown in Figure 39.
87
DMG TYPE 7 - FLEXUP
Electricity (kW)
3000
2500
2000
1500
1000
500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
FLEXUP (kW)
Imported Electricity (kW)
DMG TYPE 7 - FLEXDOWN
2500
Electricity (kW)
2000
1500
1000
500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (h)
FLEXDOWN (kW)
exported electricity (kW)
Figure 39 DMG type 7 additional operational flexibility for a spring/autumn day a) upward, b) downward
88
DMG TYPE 7 - FLEXUP
4000
Electricity (kW)
3500
3000
2500
2000
1500
1000
500
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
0
Time (kW)
FLEXUP (kW)
imported electricity (kW)
2000
1800
1600
1400
1200
1000
800
600
400
200
0
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
Electricity (kW)
DMG TYPE 7 - FLEXDOWN
Time (h)
FLEXDOWN (kW)
exported electricity (kW)
Figure 40 DMG type 7 additional operational flexibility for a summer day a) upward, b) downward
Unlike for spring and winter day, during summer day there is a lack of flexibility for the DMG
unit to be able to work as an isolated system. This can be noticed in Figure 40 in the evening
periods for both directions of flexibility. There are several possibilities to improve this:
a) Instead of single CHP unit several smaller ones could be considered. This in terms would
reduce the MSG of the CHP components of DMG unit.
b) Increasing TES size, providing more flexibility upward and downward.
89
c) Increasing EHP size.
Detailed analysis of off-grid operation of DMG units as well as defining the flexibility role of
multi-generation systems is an interesting field for future research work.
F. Discussion
Decarbonisation of all energy sectors is being steered towards integration of renewable energy
sources characterized by uncertainty and variability in production. If the integration of these
technologies is to have a positive impact on the environment and keep the system operational
cost as low as possible, new sources and concepts for provision of flexibility are needed to
maintaining the equilibrium of demand and generation.
In this Chapter an extensive overview of the flexibility issues is given, clearly recognizing
existing and potential sources of flexibility as well as their technical capabilities and their
constraints to respond to system requirements. Existing research attempts to capture and define
this flexibility either by simulating different operation scenarios or by, in addition, defining an
offline metric evaluating the flexibility of the existing or planned energy mix. However, rarely
are those metrics properly capturing relevant technical characteristics, and never are they able to
capture interactions and benefits of coupling different energy vectors. Through this review
relevant parameters for defining flexibility are extrapolated, elaborated and assigned to flexibility
sources according to their technical characteristics and constraints. In addition, the work
recognizes the potential value that interconnecting energy vectors can have on planning and
operation of future sustainable energy system, making a clear distinction from capturing
decarbonisation issues from electricity only point of view.
90
MULTI-ENERGY VIRTUAL POWER PLANT
A. The concept of Virtual Power Plant
Increasing concerns over environmental impact of the conventional fossil fuelled power plants
resulted in constant growth of renewable energy sources (RES) during the last couple of decades.
However, RES cannot yet provide levels of return on investment like fossil fuels [151] and for
this reason various incentive schemes supporting electricity produced from RES have been
introduced. In 2012 the worldwide wind power capacity reached 282 GW, having 44.6 GW of
new wind power plants installed. That is an annual growth rate of 19% [152]. At the same time
2013 was a record year for newly installed photovoltaic (PV) capacity reaching 139 GW,
installing around 39 GW in 2013 only [153]. Nevertheless, the government incentives have a
time limit after which RESs will have find a business case making the profit in the electricity
market. Exposing RES to the rigorous market environment poses a serious challenge for RES
owners, the main reason being the uncertainty of forecasted power output of RES. For instance,
wind power plants (WPPs) are inherently intermittent due to stochastic nature of wind, and PV
power plants’ output depends on solar irradiation and clouds. Thus, the risk of not meeting longterm and mid-term electricity delivery contracts is imminent. A number of concepts and
proposals, especially for wind producer bidding strategies, are reported in the literature. In [154]
a stochastic model, which minimizes the imbalance costs, is developed to generate the optimal
wind power producer bids in a short-term market. A technique which results in best offering
strategy of a wind power producer at different trading stages is presented in [155]. A suitable
trading option for wind power, based on wind power impact on market clearing prices, is
proposed in [156]. Offering strategies of interest to wind power producers are examined in [157].
The aforementioned papers do not include any kind of support technology that could improve the
wind power plant revenue. In [158] a combined strategy for bidding and operating in a power
exchange is presented. This strategy considers the combination of a wind-generation company
and a hydro-generation company. Although hydro-wind optimization is the most common
concept to cope with the uncertainty and variability of RES production [159], [160], having very
low marginal generation cost at the same time and ensuring favourable position in the market,
similar concepts are proposed with storage and flexible demand [161], [162].
91
Majority of above mentioned research deals with participation of large scale RES units in the
market, while in the context of distributed generation units (smaller units) a concept of virtual
power plant (VPP) is suggested, combining different types of renewable and non-renewable
generators with storage devices into a single market entity capable of controlling its power
output. From the point of view of any other market agent, a VPP is a unique entity, although in
reality it represents a mixture of multiple distributed energy resources (DERs) and conventional
power plants [163]. First concept of virtual power plant was described in [164], as a mix of
different generator technologies. Incorporating distributed power plants into a single legal
subject, with substantially higher installed capacity than single units have, suggests the VPP
owner is connecting its power plant to the transmission instead of the distribution grid meaning
that, from the point of view of the Transmission System Operator (TSO), a VPP is connected to
the transmission network as a single generation/consumption node using unique electricity meter.
Therefore, the VPP internal dispatch between generators inside VPP is the problem the
aggregator, and crucial in order to achieve optimal results in the both directly contracted
electricity delivery and electricity market. Economical, technical and communication issues have
been a topic of several doctoral theses [165], [166], [167]. Also through several relevant
European Commission (EC) funded projects [109], [168], [169] basic concepts and advantages of
aggregating distributed generation have been defined. Project FENIX [169], considered as the
most relevant project researching this concept, defines technical [170] and economic Virtual
Power Plant, making a distinction between goals that could drive the behaviour of distributed
units. Due to insufficient signals that would drive the operation of a technical VPP, the second
type, economic signal driven operation VPP, is a more common approach. In this context the
project recognizes two types of VPP according to the geographical characteristics and energy
needs of countries where these concepts could be implemented. The first scenario, focusing on
the example of Spain, is aimed at integrating small scale RES, usually PV and wind installed at
domestic level, and is called Southern scenario. The second VPP scenario is specific for middle
and northern Europe where CHP units are dominant due to relatively high heat demand. Similar
concept as in the specific scenario will be further elaborated in this Chapter, specifically focusing
on Multi-Energy Virtual Power Plant (MEVPP) flexibility in providing different services. Unlike
the Northern scenario in project FENIX, which brings a conclusion that VPP composed of CHP
92
units only will not gain additional profit by participating in ancillary service market, this Chapter
elaborates how and to what extent can additional flexibility be gained from MEVPP.
Each electricity market subject in the liberalized market can participate and offer their services in
different markets. For example, in case of a voluntary pool, generating companies have both open
market and bilateral contracts at their disposal. Bilateral contracts are usually concluded in the
long-term. Major reasons for bilateral contracting are price volatility and possible TSO
constraints. Each generating company decides how much of its capacity will be contracted
bilaterally in advance, and how much will be offered in the market. On the other hand, market
trading may have various time effects, ranging from the day-ahead [171] to the balancing realtime market [172]. Several papers propose simultaneous energy and market participation of VPP
using genetic algorithms [173], [174] and linear programming techniques [175], [176]. In this
Chapter only day-ahead market participation of the VPP will be considered.
B. Modelling and economic analysis of Multi-Energy Virtual Power Plant
The concept of MEVPP is envisioned as a single aggregator controlling and optimizing a number
of DMG units. Although the concept of generation unit/demand aggregation is shown in the
previous Chapter, the proposed MEVPP has its own specifics and is shown in Figure 41.
ELECTRICITY MARKET
VIRTUAL POWER PLANT
Eexp
Eexp
Eimp
Eexp
Eimp
Eimp
Eimp
Eexp
EDS
EDS
Echp
Fchp
Faux
CHP
ηe, ηe
Auxiliary boiler
ηt
Hchp
Haux
UNIT 1
Fchp
EDS
Ehp
αEchp
CHP
ηe, ηt
Electric Heat
Pump
COP
Hchp
Haux
Faux
Auxiliary boiler
ηaux
UNIT 2
ED
Hhp
C
O
N
S
U
M
E
R
S
ED, HD
(1-α)Echp
ED
C
O
N
S
U
M
E
R
S
ED, HD
(1-α)Echp
Faux
Fchp
EDS
ED
Ehp
αEchp
CHP
ηe, ηt
Hchp
Electric Heat
Pump
COP
Hhp
TES
Hs
Haux
Faux
Auxiliary boiler
ηaux
UNIT 3
C
O
N
S
U
M
E
R
S
ED, HD
ED
Fchp
… … ...
CHP
ηe, ηe
Echp
Hchp
TES
Hs
Faux
Auxiliary boiler
ηt
Haux
UNIT N
Figure 41 MEVPP concept of aggregating DMG units with local heat demand and MEVPP level electricity
demand
HD
C
O
N
S
U
M
E
R
S
ED, HD
93
The mathematical formulation of the problem is similar to those of a single unit only aggregating
them into a portfolio of units; however small modifications are needed due to locality of heat
demand and aggregated electricity demand. For this reason only equations showing equilibrium
of production and consumption will be shown again in (7.1) and (7.2):
H s (t − 1, b) + H chp (t , b) + H aux (t , b) + H ehp (t , b) − H s (t , b) =
H D (t , b)
∑ (E
chp
b∈B
E D (t )
(t , b) − Eexp (t , b) + Eimp (t , b) − Eehp (t , b)) =
(7.1)
(7.2)
The formulation of the problem is similar to a single unit problem aggregating single units of
different types, denoted with b in equations. It is clear that the MEVPP concept can be designed
of different unit types in a large range of options. For the purpose of analyses, 7 MEVPP types
are defined each composed of 60 units. For easier understanding these 7 types are shown in Table
10, each type defined by a number of different units.
Table 10 Different concepts of aggregating DMG into MEVPP
DMG TYPE 2
DMG TYPE 3
DMG TYPE 4
DMG TYPE 5
DMG TYPE 6
MEVPP 1
30
30
-
-
-
DMG TYPE
7
-
MEVPP 2
-
-
30
30
-
-
MEVPP 3
-
-
-
-
30
30
MEVPP 4
20
20
5
5
5
5
MEVPP 5
5
5
20
20
5
5
MEVPP 6
5
5
5
5
20
20
MEVPP 7
10
10
10
10
10
10
For each of the defined MEVPP a detailed sensitivity analysis is conduced, varying gas and
electricity prices ±25% and ±50% from current prices in all possible combinations. These results
are presented in Figure 42, a concept of operation costs and aggregation benefits matrix of
MEVPP.
The X axis of the Figure 42 contains electricity prices varying from -50% to +50% of todays
prices, while Y axis presents the same variation of gas prices. Each (x,y) point on Figure 42
contains a graph of operational prices of each defined MEVPP. In addition each graph presents
94
benefits of aggregating units expressed as additional operational cost savings comparing it to
uncoordinated DMG units. From the operational cost and aggregation benefits matrix several
relevant conclusions can be drawn:
•
For low electricity prices (-50% of the current electricity price) MEVPP 1, which is
composed of EHP units only, has the lowest operational cost. However, MEVPP 3
composed only of DMG type 6 and DMG type 7 units shows similar operational cost and,
in addition, it becomes the most favourable in cases with low electricity and low gas
prices (both -50% of the current prices). This again demonstrates high flexibility of DMG
6 and DMG 7 to varying market prices.
•
For high electricity prices (+50% of current electricity prices) MEVPP 1 is the least
favourable. If the gas price increases as well (both prices are +50% of todays prices) this
levels up and all MEVPP have similar operational costs. In cases when gas price is
reduced (the extreme case when electricity is 50% higher than today and gas 50% lower
than today) the lowest operational cost, as expected, have MEVPP based on CHP units MEVPP 2 and MEVPP 3.
•
In scenarios where electricity prices remain at the same level as today, it can be seen
(varying prices of gas) MEVPP 3 is the most favourable concept for all changes. Similar
can be applied to scenarios when gas prices remain as today and electricity prices are
changed in the interval [-50%, +50%].
95
MEVPP 1 operational cost
MEVPP 2 operational cost
MEVPP 3 operational cost
MEVPP 4 operational cost
MEVPP 5 operational cost
Benefits of aggregations
Figure 42 Operation costs and aggregation benefits matrix of MEVPP
MEVPP 6 operational cost
MEVPP 7 operational cost
96
•
In terms of aggregation benefits, for almost all cases, spikes for MEVPP 2 and MEVPP 4
can be noticed. For MEVPP 2 spikes can be easily explained with flexibility gain from
coupling non-flexible DMG4 and flexible DMG 5. DMG 4, when operating as a single
unit, is driven only with heat demand. When coupled with DMG 5 the excess electricity
that it produces is used to cover the electricity demand (of all units). Flexible DMG 5 in
those cases covers local heat from the storage, and CHP unit is not operating, thus saving
fuel and reducing the operational cost. In case of MEVPP 4 benefits of combining EHP
and CHP based types can be seen the best. Coordinated operation enables EHP to use
excess electricity from CHP types to produce local heat demand. The concept is in fact,
only on a larger scale, similar to DMG type 7.
•
For current electricity prices, changing only gas prices, benefits of aggregation are
ranging from 3-8% for MEVPP composed of same type units and between 8 and 15% for
MEVPP combined of different unit types.
•
For current gas prices, changing only electricity prices, benefits of aggregation are in
range of 2-8% from MEVPP composed of same type units, 7-18% for MEVPP combined
of different unit types.
•
Extreme cases from both operational cost and aggregation benefits can be noticed during
low gas prices and high electricity prices. MEVPP composed of units based on CHP have
negative operational costs and thus high aggregation benefits.
C. Future work
Presented analyses of Multi-Energy Virtual Power Plant economics are based on the assumption
that MEVPP is not a significant player in the system, meaning that the share of DMG is
insufficient for them to have an impact on creating market prices. Thus, MEVPP is a price taker.
In case MEVPP becomes a significant component of future low carbon system, there is a need for
a different approach in assessing the benefits of MEVPP. To properly define the impact MEVPP
would have on the system, an entire system model is required treating MEVPP as a price making
component. Such a comprehensive analysis would in a synthetic way define the benefits of the
DMG on a system scale. This approach is a logical direction for future research. A simplified
97
concept presenting MEVPP as a price making subject in the electricity market is shown in Figure
43.
ENERGY
NUCLEAR PP
PRICES
ENERGY
WIND PP
PRICES
PRICES
ENERGY
HYDRO PP
ENERGY
ENERGY
COAL PP
PRICES
GAS PP
PRICES
PRICES
ENERGY
ELECTRICITY MARKET
MEVPP
DEMAND
HEATING SYSTEM
Figure 43 Simplified model of price maker concept of MEVPP
98
CONCLUSION
While a great number of studies are focusing on the emerging concept of an electrical Smart
Grid, there is yet relatively little attention paid to the techno-economic and environmental
benefits that smart integration of multiple energy vectors could bring. Most analyses still consider
decoupled energy vectors and propose operational and strategic development schemes,
particularly for DH systems, independent of each other. In addition, there is no systematic
comparison of the benefits that alternative energy system options could bring. The presented
thesis introduces a unified and comprehensive techno-economic and environmental framework,
supported by a mathematical optimization model based on MILP, to compare high efficiency and
high flexibility DMG options for district energy system applications. Economic benefits on a
daily and annual operational basis have been analysed and discussed, supported by an investment
analysis for planning purposes which was tested for robustness through various sensitivity
analyses. A comprehensive environmental assessment through global (primary energy and CO2
emission reduction) and local (NOx and CO pollutants) emission analysis has also been
performed for different options. The results highlight how flexible integrated schemes with CHP
and EHP, in case supported by TES, bring cost savings as high as 50% compared to the reference
case of DH boiler. Owing to their intrinsic flexibility to respond to external conditions, such
schemes also prove extremely robust to substantial changes in energy prices and perform well
even with demanding rates of return. In addition, the environmental benefits from advanced
DMG options can also be significant, and the thesis highlights how decarbonisation of the gas
sector by just 10% could allow substantial environmental benefits for such DMG systems even
with electrical grid decarbonisation. Further benefits from flexible utilization of integrated DMG
schemes could accrue by playing in intraday balancing and ancillary services markets. The
presented research on economic and environmental benefits of flexible DMG has been published
in [149] and [177].
In addition, the thesis introduces the concept of aggregated DMG units called Multi-Energy
Virtual Power Plant. Single units cannot participate in the market or, in rare cases when they can,
their participation is limited due to small power. A concept of aggregated participation, in order
to improve market position of DMG, is presented through detailed economic analyses for a range
of possible aggregation mixes and varying prices of both gas and electricity. The results are
presented in an operational cost and aggregation benefits matrix concept, a novel approach for
99
assessing DMG benefits. While aggregation benefits vary depending on the market conditions,
the concept clearly shows the benefits of aggregating different types of DMG units as they
demonstrate high flexibility regardless of the market price scenarios.
The flexibility of DMG units primarily comes from their possibility to shift from one energy
vector to another and, by doing so, to always select the optimal input vector for provision of the
desired output. This concept is additionally emphasised by introducing multi-energy operational
flexibility metric. Alongside defining the mathematical concept for it, the metric has been tested
against the most comprehensive DMG type. The additional flexibility that can be gained is
dependent on the operational points, meaning that by providing this additional flexibility, the unit
maintains the desired outputs to the final consumer but also fulfils the contracted obligations with
the energy market. Capability of DMG units to provide additional flexibility by shifting between
different components brings benefits in terms of further reducing operational cost by providing
systems ancillary services. Parts of this work have been submitted for publication [178].
Through the thesis original scientific contributions have been presented:
1
Unified mathematical model for market and environment driven optimization of
Distributed Multi-Generation unit capable of analyzing different Multi-Generation
units behaviour in liberalized market environment and to evaluate local and global
environmental impact
2
Mathematical model of aggregated Distributed Multi-Generation units in a concept of
market driven Virtual Multi-Generation Power Plant as a new market entity in the
future low carbon energy system.
3
Methodology for evaluation of Distributed Multi-Generation units' flexibility and
definition of flexibility maps as potential capacity of Multi-Generation units for
provision of electricity systems ancillary services.
100
BIBLIOGRAPHY
[1]
P. Mancarella, Smart Multi-Energy Grids: Concepts, benefits and challenges, 2012 IEEE
Power Energy Soc. Gen. Meet. (2012) 1–2.
[2]
P. Mancarella, G. Chicco, Distributed multi-generation systems: Energy models and
analyses, Nova Science Publishers, New York, 2009.
[3]
G. Chicco, P. Mancarella, Distributed multi-generation: A comprehensive view, Renew.
Sustain. Energy Rev. 13 (2009) 535–551.
[4]
C. Yang, Hydrogen and electricity: Parallels, interactions, and convergence, Int. J.
Hydrogen Energy. 33 (2008) 1977–1994.
[5]
P. Mancarella, MES (multi-energy systems): an overview of concepts and evaluation
models, Energy. 65 (2014) 1–17.
[6]
C. Weber, N. Shah, Optimisation based design of a district energy system for an eco-town
in the United Kingdom, Energy. 36 (2011) 1292–1308.
[7]
H. Aki, T. Oyama, K. Tsuji, Analysis of energy pricing in urban energy service systems
considering a multiobjective problem of environmental and economic impact, IEEE Trans.
Power Syst. 18 (2003) 1275–1282.
[8]
V. Curti, D. Favrat, M.R. von Spakovsky, An environomic approach for the modeling and
optimization of a district heating network based on centralized and decentralized heat
pumps, cogeneration and/or gas furnace. Part II: Application, Int. J. Therm. Sci. 39 (2000)
731–741.
[9]
J.H. Horlock, Advanced Gas Turbine Cycles, Elsevier Science Ltd, Oxford, 2003.
[10] A. Martens, The energetic feasibility of CHP compared to the separate production of heat
and power, Appl. Therm. Eng. 18 (1998) 935–946.
[11] C. a. Frangopoulos, A method to determine the power to heat ratio, the cogenerated
electricity and the primary energy savings of cogeneration systems after the European
Directive, Energy. 45 (2012) 52–61.
[12] M.F. Torchio, G. Genon, A. Poggio, M. Poggio, Merging of energy and environmental
analyses for district heating systems, Energy. 34 (2009) 220–227.
[13] C. Weber, D. Favrat, Conventional and advanced CO2 based district energy systems,
Energy. 35 (2010) 5070–5081.
101
[14] P. Mancarella, C. Gan, G. Strbac, Evaluation of the impact of electric heat pumps and
distributed CHP on LV networks, in: PowerTech, 2011 IEEE, IEEE, Trondheim, 2011: pp.
1–7.
[15] H. Lund, a. N. Andersen, Optimal designs of small CHP plants in a market with
fluctuating electricity prices, Energy Convers. Manag. 46 (2005) 893–904.
[16] G. Westner, R. Madlener, Development of cogeneration in Germany: A mean-variance
portfolio analysis of individual technology’s prospects in view of the new regulatory
framework, Energy. 36 (2011) 5301–5313.
[17] E. Dotzauer, Experiences in mid-term planning of district heating systems, Energy. 28
(2003) 1545–1555.
[18] B. Rolfsman, Combined heat-and-power plants and district heating in a deregulated
electricity market, Appl. Energy. 78 (2004) 37–52.
[19] JDS Associates, Micro-CHP as an integral component of the UK ’ s energy strategy, 2011.
[20] G. Davies, The potential and Cost of District Heating Networks, 2009.
[21] H. Singh, a. Muetze, P.C. Eames, Factors influencing the uptake of heat pump technology
by the UK domestic sector, Renew. Energy. 35 (2010) 873–878.
[22] A. Fragaki, A.N. Andersen, Conditions for aggregation of CHP plants in the UK electricity
market and exploration of plant size, Appl. Energy. 88 (2011) 3930–3940.
[23] A. Fragaki, A.N. Andersen, D. Toke, Exploration of economical sizing of gas engine and
thermal store for combined heat and power plants in the UK, Energy. 33 (2008) 1659–
1670.
[24] DHC+ Technology Platform, District-Heating-Barometer, EcoHeat4eu. (2011).
[25] M. Blarke, H. Lund, Large-scale heat pumps in sustainable energy systems: System and
project perspectives, Therm. Sci. 11 (2007) 143–152.
[26] B. Mathiesen, M. Blarke, K. Hansen, D. Connolly, The role of large-scale heat pumps
for short term integration of renewable energy, 2011.
[27] B.V.M.H. Lund, Comparative analyses of seven technologies to facilitate the integration of
fluctuating renewable energy sources, IET Renew. Power Gener. 3 (2009) 190–204.
[28] P. Mancarella, G. Chicco, Distributed multi-generation systems: Energy models and
analyses, Nova Science Publishers, 2009.
102
[29] Z. Bogdan, D. Kopjar, Improvement of the cogeneration plant economy by using heat
accumulator, Energy. 31 (2006) 2285–2292.
[30] A. Gebremedhin, Introducing District Heating in a Norwegian town – Potential for
reduced Local and Global Emissions, Appl. Energy. 95 (2012) 300–304.
[31] V. Curti, M.R. von Spakovsky, D. Favrat, An environomic approach for the modeling and
optimization of a district heating network based on centralized and decentralized heat
pumps, cogeneration and/or gas furnace. Part I: Methodology, Int. J. Therm. Sci. 39 (2000)
721–730.
[32] P. Mancarella, Cogeneration systems with electric heat pumps: Energy-shifting properties
and equivalent plant modelling, Energy Convers. Manag. 50 (2009) 1991–1999.
[33] P.C. Few, M. a. Smith, J.W. Twidell, Modelling of a combined heat and power (CHP)
plant incorporating a heat pump for domestic use, Energy. 22 (1997) 651–659.
[34] P. Mancarella, G. Chicco, Demand Response from Energy Shifting in Distributed MultiGeneration, IEEE Trans. Smart Grid. In Press (2014).
[35] P. Mancarella, G. Chicco, Integrated Energy and Ancillary Services Provision in MultiEnergy Systems, in: 2013 IREP Symp. Power Syst. Dyn. Control, 2013: pp. 1–19.
[36] M.B. Blarke, Towards an intermittency-friendly energy system: Comparing electric boilers
and heat pumps in distributed cogeneration, Appl. Energy. 91 (2012) 349–365.
[37] M.B. Blarke, E. Dotzauer, Intermittency-friendly and high-efficiency cogeneration:
Operational optimisation of cogeneration with compression heat pump, flue gas heat
recovery, and intermediate cold storage, Energy. 36 (2011) 6867–6878.
[38] H.B. Piittgen, P.R. Macgregor, Optimum Scheduling Procedure For Cogenerating Small
Power Producing Facilities,, IEEE Trans. Power Syst. 4 (1989).
[39] P.R. Macgregor, H.B. Puttgen, A Spot Price Based Control Mechanism For Electric Utility
Systems With Small Power Producing Facilities, IEEE Trans. Power Syst. (1991) 683–
690.
[40] P.S.N. Rao, Combined Heat and Power Economic Dispatch: A Direct Solution, Electr.
Power Components Syst. 34 (2006) 1043–1056.
[41] M.I. Henwood, M. van Ooijen, An algorithm for combined heat and power economic
dispatch, IEEE Trans. Power Syst. 11 (1996) 1778–1784.
[42] K. Heussen, S. Koch, A. Ulbig, G. Andersson, Energy Storage in Power System
Operation : The Power Nodes Modeling Framework, in: Innov. Smart Grid Technol. Conf.
Eur. (ISGT Eur. 2010 IEEE PES, 2010: pp. 1–8.
103
[43] F. Salgado, P. Pedrero, Short-term operation planning on cogeneration systems: A survey,
Electr. Power Syst. Res. 78 (2008) 835–848.
[44] O. a. Shaneb, P.C. Taylor, G. Coates, Optimal online operation of residential μCHP
systems using linear programming, Energy Build. 44 (2012) 17–25.
[45] F. Freschi, L. Giaccone, P. Lazzeroni, M. Repetto, Economic and environmental analysis
of a trigeneration system for food-industry: A case study, Appl. Energy. 107 (2013) 157–
172.
[46] E.D. Mehleri, H. Sarimveis, N.C. Markatos, L.G. Papageorgiou, A mathematical
programming approach for optimal design of distributed energy systems at the
neighbourhood level, Energy. 44 (2012) 96–104.
[47] G.M. Kopanos, M.C. Georgiadis, E.N. Pistikopoulos, Energy production planning of a
network of micro combined heat and power generators, Appl. Energy. 102 (2013) 1522–
1534.
[48] I.E. Grossmann, S.A. Papoulias, A Structural Optimization Approach Process Syntesis-I
Utility Systems, Comput. Chem. Eng. 7 (1983).
[49] I.E. Grossmann, S.A. Papoulias, A Structural Optimization Approach Process Synthesis-II
Heat Recovery Networks, Comput. Chem. Eng. 7 (1983) 707–721.
[50] P. Voll, C. Klaffke, M. Hennen, A. Bardow, Automated superstructure-based synthesis
and optimization of distributed energy supply systems, Energy. 50 (2013) 374–388.
[51] P. Voll, M. Lampe, G. Wrobel, A. Bardow, Superstructure-free synthesis and optimization
of distributed industrial energy supply systems, Energy. 45 (2012) 424–435.
[52] P.S. Woods, G. Zdaniuk, CHP and District Heating - how efficient are these technologies?,
in: CIBSE Tech. Symp., 2011: pp. 1–17.
[53] G. Chicco, P. Mancarella, Trigeneration primary energy saving evaluation for energy
planning and policy development, Energy Policy. 35 (2007) 6132–6144.
[54] G. Chicco, P. Mancarella, A unified model for energy and environmental performance
assessment of natural gas-fueled poly-generation systems, Energy Convers. Manag. 49
(2008) 2069–2077.
[55] P. Mancarella, G. Chicco, Distributed Cogeneration : Modelling of Environmental Benefits
and Impact, in: D N Gainkar (Ed.), Distrib. Gener., InTech, 2010.
[56] P. Mancarella, G. Chicco, Global and local emission impact assessment of distributed
cogeneration systems with partial-load models, Appl. Energy. 86 (2009) 2096–2106.
104
[57] A. Canova, G. Chicco, G. Genon, P. Mancarella, Emission characterization and evaluation
of natural gas-fueled cogeneration microturbines and internal combustion engines, Energy
Convers. Manag. 49 (2008) 2900–2909.
[58] University of Strathclyde, Demand Profile Generator, (n.d.).
[59] V. Verda, E. Guelpa, A. Kona, S. Lo Russo, Reduction of primary energy needs in urban
areas trough optimal planning of district heating and heat pump installations, Energy. 48
(2012) 40–46.
[60] Elexon Ltd, Elexon, BM Reports. (2012).
[61] Department of Energy&Climate Change, Energy price statistics, (n.d.).
[62] V. Verda, F. Colella, Primary energy savings through thermal storage in district heating
networks, Energy. 36 (2011) 4278–4286.
[63] M. Blarke, Smart intermittency-friendly cogeneration: Techno-economic performance of
innovative double storage concept for integrating compression heat pumps in distributed
cogeneration, in: 6th Int. Renew. Energy Storage Conf., 2011: pp. 1–20.
[64] A. Pini Prato, F. Strobino, M. Broccardo, L. Parodi Giusino, Integrated management of
cogeneration plants and district heating networks, Appl. Energy. 97 (2012) 590–600.
[65] M. Houwing, R.R. Negenborn, B. De Schutter, Demand Response With Micro-CHP
Systems, Proc. IEEE. 99 (2011) 200–213.
[66] G. Papaefthymiou, B. Hasche, C. Nabe, Potential of Heat Pumps for Demand Side
Management and Wind Power Integration in the German Electricity Market, IEEE Trans.
Sustain. Energy. 3 (2012) 636–642.
[67] E. Carpaneto, G. Chicco, P. Mancarella, A. Russo, Cogeneration planning under
uncertainty Part I: Multiple time frame approach, Appl. Energy. 88 (2011) 1059–1067.
[68] E. Carpaneto, G. Chicco, P. Mancarella, A. Russo, Cogeneration planning under
uncertainty. Part II: Decision theory-based assessment of planning alternatives, Appl.
Energy. 88 (2011) 1075–1083.
[69] K. Siler-Evans, M.G. Morgan, I.L. Azevedo, Distributed cogeneration for commercial
buildings: Can we make the economics work?, Energy Policy. 42 (2012) 580–590.
[70] Department of Energy&Climate Change, Digest of United Kingdom Energy Statistics
2013, 2013.
[71] A.D. Hawkes, M.A. Leach, On policy instruments for support of micro combined heat and
power, Energy Policy. 36 (2008) 2973–2982.
105
[72] A.D. Hawkes, Estimating marginal CO2 emissions rates for national electricity systems,
Energy Policy. 38 (2010) 5977–5987.
[73] DECC, Fuel Mix Disclosure data table, (2012).
[74] A.F.M. for Transport, Feeding biogas into the Austrian natural gas grid, 2006.
[75] T.K. Jensen, Biogas to the natural gas grid, 2010.
[76] National Grid, The potential for Renewable Gas in the UK, 2009.
[77] Department of Energy & Clima Change, Biomethane into the Gas Network: A Guide for
Producers, 2009.
[78] Anaerobic Digestion, Renewable Heat Incentive, (n.d.).
[79] Environmental Protection Agency Combined Heat and Power Partnership, Technology
Characterization: Reciprocating Engines, Washington, DC, 2008.
[80] T. Capuder, P. Mancarella, Techno-economic and Environmental Modelling and
Optimization of Flexible Distributed Multi-Generation Options, Energy. (2014).
[81] M.B. Blarke, H. Lund, The effectiveness of storage and relocation options in renewable
energy systems, Renew. Energy. 33 (2008) 1499–1507.
[82] H. Chandler, A rewiev of measure to increase the flexibility of power systems in response
to larger shares of variable output renewables, in: 3rd Int. Conf. Integr. Renew. Distrib.
Energy Resour., Nice, France, 2008.
[83] D. Lew, Finding Flexibility in the Existing Grid, 2011.
[84] B. Kirby, M. Milligan, A Method and Case Study for Estimating the Ramping Capability
of a Control Area or Balancing Authority and Implications for Moderate or High Wind
Penetration, 2005.
[85] M. Milligan, B. Kirby, Impact of Balancing Areas Size, Obligation Sharing , and Ramping
Capability on Wind Integration, 2007.
[86] Market Quality &Renewable Integration, 2013 Ancillary Service Scarcity Event Report,
2014.
[87] N. Navid, G. Rosenwald, Market Solutions for Managing Ramp Flexibility With High
Penetration of Renewable Resource, IEEE Trans. Sustain. Energy. 3 (2012) 784–790.
[88] B.T. Ackermann, G. Ancell, L.D. Borup, P.B. Eriksen, B. Ernst, F. Groome, et al., Where
the Wind Blows, IEEE Power Energy Mag. (2009) 65–75.
106
[89] RWE, The Need for Smart Megawatts Power Generation in Europe – Facts & Trends
December 2009, 2009.
[90] M. Nicolosi, Wind power integration and power system flexibility–An empirical analysis
of extreme events in Germany under the new negative price regime, Energy Policy. 38
(2010) 7257–7268.
[91] G. Keane, Options for Low-Carbon Power Sector Flexibility to 2050, 2010.
[92] K. Hedegaard, P. Meibom, Wind power impacts and electricity storage – A time scale
perspective, Renew. Energy. 37 (2012) 318–324.
[93] P. Denholm, E. Ela, B. Kirby, M. Milligan, The Role of Energy Storage with Renewable
Electricity Generation, 2010.
[94] P. Denholm, M. Hand, Grid flexibility and storage required to achieve very high
penetration of variable renewable electricity, Energy Policy. 39 (2011) 1817–1830.
[95] A. Tuohy, M. O’Malley, Impact of pumped storage on power systems with increasing
wind penetration, in: 2009 IEEE Power Energy Soc. Gen. Meet., Ieee, 2009: pp. 1–8.
[96] G. Strbac, A. Shakoor, M. Black, D. Pudjianto, T. Bopp, Impact of wind generation on the
operation and development of the UK electricity systems, Electr. Power Syst. Res. 77
(2007) 1214–1227.
[97] P.D. Brown, J.A.P. Lopes, M.A. Matos, Optimization of Pumped Storage Capacity in an
Isolated Power System With Large Renewable Penetration, IEEE Trans. Power Syst. 23
(2008) 523–531.
[98] A.A. Akhil, G. Huff, A.B. Currier, B.C. Kaun, D.M. Rastler, S.B. Chen, et al., DOE /
EPRI 2013 Electricity Storage Handbook in Collaboration with NRECA, 2013.
[99] F. Oldewurtel, T. Borsche, M. Bucher, P. Fortenbacher, M. Gonz, T. Haring, et al., A
Framework for and Assessment of Demand Response and Energy Storage in Power
Systems a, in: 2013 IREP Symp. Power Syst. Dyn. Control, Rethymnon, 2013: pp. 1–24.
[100] B. Hartmann, A. Dán, Methodologies for Storage Size Determination for the Integration of
Wind Power, IEEE Trans. Sustain. Energy. 5 (2014) 182–189.
[101] Y.M. Atwa, E.F. El-Saadany, Optimal Allocation of ESS in Distribution Systems With a
High Penetration of Wind Energy, IEEE Trans. Power Syst. 25 (2010) 1815–1822.
[102] H. Le, T. Nguyen, Sizing energy storage systems for wind power firming: An analytical
approach and a cost-benefit analysis, in: IEEE PES Gen. Meet., Pittsburg, 2008: pp. 1–8.
107
[103] A. Oudalov, D. Chartouni, C. Ohler, Optimizing a Battery Energy Storage System for
Primary Frequency Control, IEEE Trans. Power Syst. 22 (2007) 1259–1266.
[104] X. He, E. Delarue, W. D’haeseleer, J.-M. Glachant, A novel business model for
aggregating the values of electricity storage, Energy Policy. 39 (2011) 1575–1585.
[105] I. Pierre, F. Bauer, R. Blasko, N. Dahlback, M. Dumpelmann, K. Kainurinne, et al.,
Flexible generation : Backing up renewables, 2011.
[106] Flexibility Requirements and Potential Metrics for Variable Generation: Implications for
System Planning Studies, 2010.
[107] N.A.E.R. Corporation, Flexibility requirements and Potential Metrics for Variable
Generation: Implications for System Planning Studies, 2010.
[108] K. Engblom, Flexible Power Generation Enabling Wide Integration of RES, 2012.
[109] A. Birke, S. Borresen, K. Edlund, K. Ilskjaer, I.K. Pedersen, N.D. Rasmussen, et al.,
Providing flexibility with a virtual power plant, 2012.
[110] F.C. Schweppe, R.D. Tabors, J.L. Kirtley, H.R. Outhred, F.H. Pickel, A.J. Cox,
Homeostatic Utility Control, IEEE Trans. Power Appar. Syst. 75 (1980) 1151–1163.
[111] A. Molina-garcía, F. Bouffard, D.S. Kirschen, Decentralized Demand-Side Contribution to
Primary Frequency Control, IEEE Trans. Power Syst. 26 (2011) 411–419.
[112] S. Koch, F.S. Barcenas, G. Andersson, Using Controllable Thermal Household Appliances
for Wind Forecast Error Reduction, in: IFAC Conf. Control Methodol. Technol. Energy
Effic., Vilamoura, 2010: pp. 1–4.
[113] E. Karangelos, F. Bouffard, Towards Full Integration of Demand-Side Resources in Joint
Forward Energy/Reserve Electricity Markets, IEEE Trans. Power Syst. 27 (2012) 280–
289.
[114] B. Biegel, L.H. Hansen, J. Stoustrup, P. Andersen, S. Harbo, Value of flexible
consumption in the electricity markets, Energy. 66 (2014) 354–362.
[115] D. Wang, S. Parkinson, W. Miao, H. Jia, C. Crawford, N. Djilali, Hierarchical market
integration of responsive loads as spinning reserve, Appl. Energy. 104 (2013) 229–238.
[116] B. Biegel, M. Westenholz, L.H. Hansen, J. Stoustrup, P. Andersen, S. Harbo, Integration
of flexible consumers in the ancillary service markets, Energy. (2014) 1–11.
[117] O. Ma, N. Alkadi, P. Cappers, P. Denholm, J. Dudley, S. Goli, et al., Demand Response
for Ancillary Services, IEEE Trans. Smart Grid. 4 (2013) 1988–1995.
108
[118] D.S. Callaway, I. a Hiskens, Achieving Controllability of Electric Loads, Proc. IEEE. 99
(2011) 184–199.
[119] B. Stigson, Mobility 2030: Meeting the challenges to sustainability, 2004.
[120] D. Papadaskalopoulos, G. Strbac, P. Mancarella, M. Aunedi, V. Stanojevic, Decentralized
Participation of Flexible Demand in Electricity Markets—Part II: Application With
Electric Vehicles and Heat Pump Systems, IEEE Trans. Power Syst. 28 (2013) 3667–3674.
[121] Y. Mu, S. Member, J. Wu, J. Ekanayake, S. Member, Primary Frequency Reponse From
Electric Vehicles in the Great Britain Power System, IEEE Trans. Smart Grid. 4 (2013)
1142–1150.
[122] W. Kempton, J. Tomić, Vehicle-to-grid power implementation: From stabilizing the grid
to supporting large-scale renewable energy, J. Power Sources. 144 (2005) 280–294.
[123] J. Pecas Lopes, F.J. Soares, P.M.R. Almeida, Integration of Electric Vehicles in the
Electric Power System, Proc. IEEE. 99 (2011).
[124] E. Sortomme, M. a. El-Sharkawi, Optimal Scheduling of Vehicle-to-Grid Energy and
Ancillary Services, IEEE Trans. Smart Grid. 3 (2012) 351–359.
[125] A. Shortt, M.O. Malley, Quantifying the Long-Term Impact of Electric Vehicles on the
Generation Portfolio, IEEE Trans. Smart Grid. 5 (2014) 71–83.
[126] M. a. Ortega-Vazquez, F. Bouffard, V. Silva, Electric Vehicle Aggregator/System
Operator Coordination for Charging Scheduling and Services Procurement, IEEE Trans.
Power Syst. 28 (2013) 1806–1815.
[127] E. Lannoye, D. Flynn, M.O. Malley, Power System Flexibility Assessment - State of the
Art, in: Power Energy Soc. Gen. Meet. 2012 IEEE, 2012.
[128] H. Chavez, R. Baldick, Inertia and Governor Ramp Rate Constrained Economic Dispatch
to Assess Primary Frequency Response Adequacy, in: Int. Conf. Renew. Energies Power
Qual., 2012: pp. 2–7.
[129] E. Ela, V. Gevorgian, A. Tuohy, B. Kirby, S. Member, M. Milligan, et al., Market Designs
for the Primary Frequency Response Ancillary Service — Part I : Motivation and Design,
IEEE Trans. Power Syst. 29 (2014) 421–431.
[130] E. Ela, V. Gevorgian, A. Tuohy, B. Kirby, S. Member, M. Milligan, et al., Market Designs
for the Primary Frequency Response Ancillary Service — Part II : Case Studies, IEEE
Trans. Power Syst. 29 (2014) 432–440.
[131] Y. V Makarov, C. Loutan, J. Ma, P. De Mello, Operational Impacts of Wind Generation
on California Power Systems, IEEE Trans. Power Syst. 24 (2009) 1039–1050.
109
[132] Y. V Makarov, P. V Etingov, J. Ma, Incorporating Uncertainty of Wind Power Generation
Forecast Into Power System Operation , Dispatch , and Unit Commitment Procedures,
IEEE Trans. Sustain. Energy. 2 (2011) 433–442.
[133] Y. V Makarov, P. Du, M.A. Pai, B. Mcmanus, Calculating Individual Resources
Variability and Uncertainty Factors Based on Their Contributions to the Overall System
Balancing Needs, IEEE Trans. Smart Grid. 5 (2014) 323–331.
[134] C. Wang, J.D. McCalley, Impact of wind power on control performance standards, Int. J.
Electr. Power Energy Syst. 47 (2013) 225–234.
[135] H. Chávez, R. Baldick, J. Matevosyan, CPS1 Compliance-Constrained AGC Gain
Determination for a Single-Balancing Authority, IEEE Trans. Power Syst. 29 (2014)
1481–1488.
[136] E. Lannoye, M. Milligan, J. Adams, A. Tuohy, H. Chandler, D. Flynn, et al., Integration of
Variable Generation : Capacity Value and Evaluation of Flexibility, in: Power Energy Soc.
Gen. Meet. 2010 IEEE, 2009: pp. 1–6.
[137] E. Lannoye, D. Flynn, M.O. Malley, Evaluation of Power System Flexibility, IEEE Trans.
Power Syst. 27 (2012) 922–931.
[138] E. Lannoye, D. Flynn, M. O’Malley, Assessment of power system flexibility: A high-level
approach, in: 2012 IEEE Power Energy Soc. Gen. Meet., Ieee, 2012: pp. 1–8.
[139] N. Menemenlis, M. Huneault, A. Robitaille, Thoughts on power system flexibility
quantification for the short-term horizon, in: 2011 IEEE Power Energy Soc. Gen. Meet.,
Ieee, 2011: pp. 1–8.
[140] J. Ma, V. Silva, R. Belhomme, D.S. Kirschen, L.F. Ochoa, Evaluating and Planning
Flexibility in Sustainable Power Systems, IEEE Trans. Sustain. Energy. 4 (2013) 200–209.
[141] D.S. Kirschen, J. Ma, V. Silva, R. Belhomme, Optimizing the flexibility of a portfolio of
generating plants to deal with wind generation, 2011 IEEE Power Energy Soc. Gen. Meet.
(2011) 1–7.
[142] R. Loisel, Power system flexibility with electricity storage technologies: A technical–
economic assessment of a large-scale storage facility, Int. J. Electr. Power Energy Syst. 42
(2012) 542–552.
[143] D. Mejía-giraldo, J.D. Mccalley, Maximizing Future Flexibility in Electric Generation
Portfolios, IEEE Trans. Power Syst. 29 (2014) 279–288.
[144] M.B. Blarke, The missing link in sustainable energy: Techno-economic consequences of
large-scale heat pumps in distributed generation in favour of a domestic integration
strategy for sustainable energy, 2008.
110
[145] T. Krause, G. Andersson, K. Fröhlich, A. Vaccaro, Multiple-Energy Carriers : Modeling of
Production , Delivery , and Consumption, Proc. IEEE. 99 (2010) 15–27.
[146] N. Keyaerts, E. Delarue, Y. Rombauts, W. D’haeseleer, Impact of unpredictable
renewables on gas-balancing design in Europe, Appl. Energy. 119 (2014) 266–277.
[147] M. Chaudry, M. Qadrdan, J. Wu, N. Jenkins, Economic impacts of wind generation
variability on gas network operation, in: 8th BIEE Acad. Conf., 2010: pp. 1–5.
[148] T. Nuytten, B. Claessens, K. Paredis, J. Van Bael, D. Six, Flexibility of a combined heat
and power system with thermal energy storage for district heating, Appl. Energy. 104
(2013) 583–591.
[149] T. Capuder, P. Mancarella, Techno-economic and Environmental Modelling and
Optimization of Flexible Distributed Multi-Generation Options, Energx. Accepted f
(2014).
[150] P. Mancarella, G. Chicco, Demand Response from Energy Shifting in Distributed MultiGeneration, IEEE Trans. Smart Grid. 4 (2013) 1928–1938.
[151] A. Cavallo, Controllable and affordable utility-scale electricity from intermittent wind
resources and compressed air energy storage (CAES), Energy. 32 (2007) 120–127.
[152] WWEA, World Wind Energy Report 2012, Bonn, 2012.
[153] European Photovoltaic Industry Association, Global Market Outlook 2014, 2013.
[154] J. Matevosyan, L. Söder, Minimization of Imbalance Cost Trading Wind Power on the
Short-Term Power Market, IEEE Trans. Power Syst. 21 (2006) 1396–1404.
[155] J.M. Morales, A.J. Conejo, J. Pérez-ruiz, Short-Term Trading for a Wind Power Producer,
IEEE Trans. Power Syst. 25 (2010) 554–564.
[156] S.N. Singh, I. Erlich, Strategies for Wind Power Trading in Competitive Electricity
Markets, IEEE Trans. Energy Convers. 23 (2008) 249–256.
[157] M. Rahimiyan, J.M. Morales, A.J. Conejo, Evaluating alternative offering strategies for
wind producers in a pool, Appl. Energy. 88 (2011) 4918–4926.
[158] J.M. Angarita, J.G. Usaola, Combining hydro-generation and wind energy, Electr. Power
Syst. Res. 77 (2007) 393–400.
[159] E.D. Castronuovo, J. a. P. Lopes, Optimal operation and hydro storage sizing of a wind–
hydro power plant, Int. J. Electr. Power Energy Syst. 26 (2004) 771–778.
111
[160] Á.J. Duque, E.D. Castronuovo, I. Sánchez, J. Usaola, Optimal operation of a pumpedstorage hydro plant that compensates the imbalances of a wind power producer, Electr.
Power Syst. Res. 81 (2011) 1767–1777.
[161] A. Schroeder, Modeling storage and demand management in power distribution grids,
Appl. Energy. 88 (2011) 4700–4712.
[162] C. De Jonghe, E. Delarue, R. Belmans, W. D’haeseleer, Determining optimal electricity
technology mix with high level of wind power penetration, Appl. Energy. 88 (2011) 2231–
2238.
[163] K. Kok, Short-term Economics of Virtual Power Plants, in: 20th Int. Conf. Electr. Distrib.,
Prague, 2009: pp. 8–11.
[164] S. Awerbuch, A. Preston, The Virtual Utility, Springer, 1997.
[165] P. Landsbergen, Feasibility , beneficiality, and institutional compatibility of a micro-CHP
virtual power plant in the Netherlands, Delft University of Technology, 2009.
[166] S. You, Developing Virtual Power Plant for Optimized Distributed Energy Resources
Operation and Integration, Technical University of Denmark, 2010.
[167] G. Newman, Characterisation of Virtual Power Plants, University of Manchester, 2010.
[168] T. Erge, C. Sauer, MASSIG Market Access for Smaller Size Intelligent Electricity
Generation, Freiburg, 2010.
[169] A. van der Welle, C. Kolokathis, J. Jansen, C. Madina, A. Diaz, FENIX ’...a step towards
the future of electricity networks’, Le Petten, 2009.
[170] D. Pudjianto, C. Ramsay, G. Strbac, Virtual power plant and system integration of
distributed energy resources, IET Renew. Power Gener. 1 (2007) 10–16.
[171] M. Peik-Herfeh, H. Seifi, M.K. Sheikh-El-Eslami, Decision making of a virtual power
plant under uncertainties for bidding in a day-ahead market using point estimate method,
Int. J. Electr. Power Energy Syst. 44 (2013) 88–98.
[172] C. Klessmann, C. Nabe, K. Burges, Pros and cons of exposing renewables to electricity
market risks—A comparison of the market integration approaches in Germany, Spain, and
the UK, Energy Policy. 36 (2008) 3646–3661.
[173] E. Mashhour, S.M. Moghaddas-tafreshi, Bidding Strategy of Virtual Power Plant for
Participating in Energy and Spinning Reserve Markets — Part I : Problem Formulation,
IEEE Trans. Power Syst. 26 (2011) 949–956.
112
[174] E. Mashhour, S.M. Moghaddas-Tafreshi, Bidding Strategy of Virtual Power Plant for
Participating in Energy and Spinning Reserve Markets—Part II: Numerical Analysis,
IEEE Trans. Power Syst. 26 (2011) 957–964.
[175] H. Pandžić, J.M. Morales, A.J. Conejo, I. Kuzle, Offering model for a virtual power plant
based on stochastic programming, Appl. Energy. 105 (2013) 282–292.
[176] P. Faria, T. Soares, Z. Vale, H. Morais, Distributed generation and demand response
dispatch for a virtual power player energy and reserve provision, Renew. Energy. 66
(2014) 686–695.
[177] T. Capuder, P. Mancarella, Modelling and Assessment of the Techno-economic and
Environmental Performance of Flexible Multi-Generation Systems, in: 18th Power Syst.
Comput. Conf., Wroclaw, 2014: pp. 1–7.
[178] T. Capuder, P. Mancarella, Flexibility - a Key Element in Operation and Planning of Low
Carbon Energy Systems, IEEE Syst. J. submitted (2014) 1–9.
[179] R. Yokoyama, T. Wakui, J. Kamakari, K. Takemura, Performance analysis of a CO2 heat
pump water heating system under a daily change in a standardized demand, Energy. 35
(2010) 718–728.
113
LIST OF FIGURES
Figure 1 Energy flow layout of a) DMG type 1, b) DMG type 2, c) DMG type 3, d) DMG type 4, e) DMG type 5, f)
DMG type 6, g) DMG type 7 ....................................................................................................................................... 11
Figure 2 DMG types creation flow diagram ................................................................................................................ 13
Figure 3 Demand profiles and electricity prices for a specific winter day ................................................................... 21
Figure 4 Daily operation of DMG type 4 for a typical winter day – electricity and heat ............................................. 22
Figure 5 Daily operation of DMG type 5 for a typical winter day – electricity and heat ............................................. 24
Figure 6 Daily operation of DMG type 6 for a typical winter day – electricity and heat ............................................. 25
Figure 7 Daily operation of DMG type 7 for a typical winter day – electricity and heat ............................................. 27
Figure 8 Demand profiles and electricity prices for a specific spring/autumn day ...................................................... 28
Figure 9 Daily operation of DMG type 6 for a typical sprig/autumn day – electricity and heat .................................. 29
Figure 10 Daily operation of DMG type 7 for a typical sprig/autumn day – electricity and heat ................................ 31
Figure 11 Demand profiles and electricity prices for a specific summer day .............................................................. 32
Figure 12 Daily operation of DMG type 6 for a typical summer day – electricity and heat ........................................ 33
Figure 13 Daily operation of DMG type 7 for a typical summer day – electricity and heat ........................................ 34
Figure 14 Results for sensitivity analyses for DMG type 5 ......................................................................................... 38
Figure 15 Results for sensitivity analyses for DMG type 6 ......................................................................................... 39
Figure 16 Results for sensitivity analyses for DMG type 7 ......................................................................................... 40
Figure 17 NPV dependency on discount rate for different DMG types ....................................................................... 41
Figure 18 Sensitivity analyses for DMG types 5, 6 and 7 ............................................................................................ 42
Figure 19 Half hourly Average Emission Factor for UK in 2011 and 2012 ................................................................ 46
Figure 20 DMG options carbon footprint .................................................................................................................... 47
Figure 21 Results for sensitivity analyses for DMG type 5 ......................................................................................... 53
Figure 22 Results for sensitivity analyses for DMG type 6 ......................................................................................... 54
Figure 23 Results for sensitivity analyses for DMG type 7 ......................................................................................... 55
Figure 24 NPV difference in sensitivity analyses for DMG type 5.............................................................................. 55
Figure 25 NPV difference in sensitivity analyses for DMG type 6.............................................................................. 56
Figure 26 NPV difference in sensitivity analyses for DMG type 7.............................................................................. 56
Figure 27 NPV dependency on discount rate for different DMG types ....................................................................... 57
Figure 28 Sensitivity analyses for DMG types 5, 6 and 7 ............................................................................................ 58
Figure 29 Modelling difference sensitivity analyses for DMG types 5, 6 and 7 .......................................................... 58
Figure 30 DMG options carbon footprint .................................................................................................................... 59
Figure 31 DMG options carbon footprint modelling differences ................................................................................. 59
Figure 32 Flexibility requirements change with increasing share of RES ................................................................... 63
Figure 33 Flexible and inflexible sources .................................................................................................................... 66
114
Figure 34 Hierarchical control of DR and VPP ........................................................................................................... 71
Figure 35 Flexibility from energy vector interaction ................................................................................................... 78
Figure 36 Current and future flexible energy systems ................................................................................................. 79
Figure 37 Flexibility provision a) optimal operation points, b) upward flexibility requirement, c) downward
flexibility requirement.................................................................................................................................................. 81
Figure 38 DMG type 7 additional operational flexibility for a winter day a) upward, b) downward .......................... 86
Figure 39 DMG type 7 additional operational flexibility for a spring/autumn day a) upward, b) downward .............. 87
Figure 40 DMG type 7 additional operational flexibility for a summer day a) upward, b) downward ........................ 88
Figure 41 MEVPP concept of aggregating DMG units with local heat demand and MEVPP level electricity demand
..................................................................................................................................................................................... 92
Figure 42 Operation costs and aggregation benefits matrix of MEVPP ...................................................................... 95
Figure 43 Simplified model of price maker concept of MEVPP ................................................................................. 97
115
LIST OF TABLES
Table 1 Optimal sizes and relevant efficiencies of units for different DMG DG schemes .......................................... 20
Table 2 Annual operational costs of different DMG types .......................................................................................... 35
Table 3 Input data for the investment analyses ............................................................................................................ 37
Table 4 Multi-generation primary energy savings for different DMG options ............................................................ 45
Table 5 Local emissions assessments for different DMG options ............................................................................... 49
Table 6 Annual operational costs of different dmg types............................................................................................. 51
Table 7 Optimal sizes and relevant efficiencies of units for different DMG DH schemes .......................................... 52
Table 8 Modelling differences in primary energy savings for different DMG options ................................................ 60
Table 9 Local emissions assessments for different DMG options ............................................................................... 60
Table 10 Different concepts of aggregating DMG into MEVPP ................................................................................. 93
116
NOMENCLATURE
Acronyms and Indices
ACE
Area Control Errors
AEF
Average Emission Factor
BA
Balancing Area
BRP
Balancing Responsible Party
CAES
Compressed Air Energy Storage
CHP
Combined Heat and Power
CO
Carbon Oxide
CO2
Carbon Dioxide
CPS1, CPS2
Control Performance Standard
DA
Day Ahead (Market)
DH
District Heating
DMG
Distributed Multi-Generation
DR
Demand Response
EHP
Electric Heat Pump
ELCC
Effective Load Carrying Capability
EV
Electric Vehicle
FF
Flexibility Factor
IREE
Insufficient Ramping Resource Expectation
LOLE
Loss od Load Expectation
LSP
Local Separate Production
MEVPP
Multi-Energy Virtual Power Plant
MILP
Mixed Integer Linear Programming
MSG
Minimum Stable Generation
NERC
North American Electric Reliability Corporation
NOx
Nitrogen Oxides
POP
Preferable Point of Operation
PPES
Poligeneration Primary Energy Saving
117
PCO2ER
Poligeneration CO2 Emission Reduction
RES
Renewable Energy Sources
SP
Separate Production
TSO
Transmission System Operator
UC
Unit Commitment
V2G
Vehicle-to-Grid
VPP
Virtual Power Plant
D
Demand
e
Electricity
th
Thermal
aux
Auxiliary
s
Thermal energy storage
1,2,3,4,5,6,7
Type of Distributed Multi-Generation option
Input values
ŋe
Electric efficiency of a CHP unit
ŋth
Thermal efficiency of a CHP unit
α
Shifting factor
COP
Coefficient of performance
ED
Electricity demand
HD
Heat demand
Echp_min
Minimum stable generation of CHP unit
ramp
Ramp rate value of the CHP unit
Decision variables
Echp
Electricity produced by CHP unit
Hchp
Heat produced by a CHP unit
Ichp
Binary variable indicating if the CHP unit is on/off
Ion, Ioff
Binary variables indicating turning on or off of the CHP unit
Hs
State of charge of thermal energy storage
Haux
Heat produced by auxiliary boiler
118
Hhp
Heat produced by EHP
Ehp
Electricity needed by EHP for heat production
Ehp_chp
Electricity produced by CHP and used by EHP for heat production
Ehp_g
Electricity from the electric grid and used by EHP for heat production
119
BIOGRAPHY
Tomislav Capuder was born in 1983 in Zagreb. In 2008 he graduated from the Faculty of
Electrical Engineering and Computing, Department of Energy and Power Systems, obtaining the
title of Master of Electrical Engineering. In 2014 he received doctoral degree from the same
University. In 2012 and 2013 he was a visiting researcher at the University of Manchester with
Dr. Pierluigi Mancarella working on Multi Energy Systems planning and modelling.
His areas of expertise include: Energy Systems Planning and Modelling, Integrated
infrastructures, Distributed Energy Systems, Energy markets, Environmental Protection in Power
System. Tomislav Capuder has published 6 journal papers and 16 international conference
papers. He participated in over 50 industry sponsored projects related to various topics of power
system planning and operation.
Published papers:
•
Capuder, Tomislav; Mancarella, Pierluigi. Techno-economic and Environmental
Modelling and Optimization of Flexible Distributed Multi-Generation Options,
Energy (Oxford). 71 (2014); 516-533
•
Capuder, Tomislav; Mancarella, Pierluigi. Modelling and Assessment of the Technoeconomic and Environmental Performance of Flexible Multi-Generation Systems,
18th Power Systems Computation Conference, Wroclaw, Poland, Aug. 18-22 2014.
•
Kuzle, Igor; Havelka, Juraj; Pandžić, Hrvoje; Capuder, Tomislav. Hands-on
Laboratory Course for Future Power System Experts, IEEE Transactions on Power
Systems, 29 (2014) , 4; 1963-1971.
•
Pandžić, Hrvoje; Kuzle, Igor; Capuder, Tomislav. Virtual power plant mid-term
dispatch optimization, Applied Energy, vol. 101, no. 1, Jan. 2013, pp. 134-141.
•
Capuder, Tomislav; Pandžić, Hrvoje; Kuzle, Igor; Škrlec, Davor. Specifics of
integration of wind power plants into the Croatian transmission network, Applied
Energy, vol. 101, no. 1, Jan. 2013, pp. 142-150.
•
Sučić, Stjepan; Dragičević, Tomislav; Capuder, Tomislav; Delimar, Marko.
Economic dispatch of virtual power plants in an event-driven service-oriented
framework using standards-based communications, Electric power systems research,
vol. 81, no. 12, Dec. 2011, pp. 2108-2119,
120
•
Tomislav Capuder, Matija Zidar, Davor Škrlec. Evolutionary algorithm with fuzzy
numbers for planning active distribution network, Electrical engineering, vol. 94, no.
3, pp. 135-145, Sep. 2012.
•
Kuzle, Igor; Holjevac, Ninoslav; Capuder, Tomislav. Model Predictive Control for
Scheduling of Flexible Microgrid Systems // Conference on Sustainable Development
of Energy, Water and Environment Systems, Venice-Istanbul, 2014
•
Zidar, Matija; Capuder, Tomislav; Georgilakis, Pavlos; Škrlec, Davor. Convex AC
Optimal Power Flow Method for Definition of Size and Location of Battery Storage
Systems in the Distribution Grid // Conference on Sustainable Development of
Energy, Water and Environment Systems, Venice-Istanbul, 2014
•
Capuder, Tomislav; Pandžić, Hrvoje; Bošnjak, Darjan; Kuzle, Igor; Škrlec, Davor.
Specifics of Integration of Wind Power Plants into the Croatian Transmission
Network// Conference on Sustainable Development of Energy, Water and
Environment Systems, Dubrovnik, Croatia, 2011
•
Capuder, Tomislav; Periša, Ivan; Hrkec, Dinko; Zidar, Matija; Tomiša, Tomislav;
Škrlec, Davor. Integration of Power Quality Monitoring System in Croatian
Distribution Network // 21st International Conference on Electricity Distribution
(CIRED), Frankfurt, Germany, 2011
•
Dragičević, Tomislav; Capuder, Tomislav; Jelavić, Mate; Škrlec, Davor. Modeling
and Simulation of Isolated DC Microgrids Supplied by Renewable Energy Resources
// Conference on Sustainable Development of Energy, Water and Environment
Systems, Dubrovnik, Croatia, 2011
•
Zidar, Matija; Capuder, Tomislav; Dragičević, Tomislav; Škrlec, Davor. Planning of
the Distribution Network Ogulin Using Optimization Tool CADDiN // CIRED 21st
International Conference on Electricity Distribution, Frankfurt, Germany, 2011
•
Capuder, Tomislav; Pandžić, Hrvoje; Bošnjak, Darjan; Kuzle, Igor; Škrlec,
Dubravka. Analysis of Incentives Approach to Renewable Sources and Cogeneration
// Proceedings of the 2010 7th International Conference on the European Energy
Market, Madrid, Spain 2010
121
•
Capuder, Tomislav; Periša, Ivan; Škrlec, Davor. Timeframe for strategy of
modernizing Croatian distribution network SCADA // IEEE Power&Energy Society
General Meeting Proceedings. Minneapolis, Minnesota, USA, 2010
•
Capuder, Tomislav; Periša, Ivan; Škrlec, Davor. Smart Vision of the Croatian
Distribution System // Smart Grids and E-Mobility, Brussels, Belgium, 2010
•
Capuder, Tomislav; Zidar, Matija; Škrlec, Davor. Distributed Generation Integration
Using Evolutionary Algorithm with Fuzzy Numbers // Proceedings of 3rd Power
Optimization and Control Conference, Gold Coast, Australia, 2010
•
Carević, Sanela; Capuder, Tomislav; Delimar, Marko. Applications of Clustering
Algorithms in Long-term Load Forecasting // Proceedings of 1st IEEE Energycon
Conference, Manama, Bahrain, 2010
•
Grganić, Hrvoje; Capuder, Tomislav; Delimar, Marko. Croatian Electric Power
System Modeling for Stability Analysis // Proceedings of 1st IEEE Energycon
Conference, Manama, Bahrain, 2010
•
Škrlec, Davor; Kuzle, Igor; Delimar, Marko; Bošnjak, Darjan; Capuder, Tomislav;
Pandžić, Hrvoje. Large Scale Wind Power Plants Integration into the Croatian Power
System // Proceedings of EPE PEMC. Ohrid, Macedonia, 2010
•
Capuder, Tomislav; Lugarić, Luka; Brekalo-Štrbić, Jurica; Krajcar, Slavko.
Optimizing the tramway power system in Zagreb // 5th IEEE Vehicle Power and
Propulsion Conference Proceedings. Dearborn, Michigan, USA 2009
•
Kuzle, Igor; Škrlec, Davor; Bošnjak, Darjan; Capuder, Tomislav; Pandžić Hrvoje.
Connection of a DFIG-based Wind Farm to the Transmission Network // Ee 2009.
Novi Sad, Serbia, 2009
122
ŽIVOTOPIS
Tomislav Capuder rođen je u Zagrebu 1983. Diplomirao je na Fakultetu elektrotehnike i
računarstva, Sveučilišta u Zagrebu, 2008. godine na smjeru elektroenergetika. Doktorirao je 2014
na Fakultetu elektrotehnike i računarstva, Sveučilišta u Zagrebu. U 2012. i 2013. boravio je na
stručnom usavršavanju na Sveučilištu u Manchesteru gdje je sa Dr. Pierluigijem Mancarellom
radio na temi modeliranja, upravljanja i planiranja višegeneracijskih sustava.
Tomislav Capuder objavio je 6 radova u međunarodnim časopisima A kategorije te 16 radova na
međunarodnim konferencijama, te je urednik dva zbornika radova. Sudjelovao je kao suradnik na
više od 50 projekata i studija sa industrijom i to iz raznih područja planiranja i vođenja
elektroenergetskog sustava.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertisement