Applied Energy 264 (2020) 114707

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Critical evaluation of European balancing markets to enable the T

participation of Demand Aggregators
Mattia Barberoa, , Cristina Corcheroa,b, Lluc Canals Casalsa,b, Lucia Igualadaa, F.-Javier Herediab
⁎

a
Catalonia Institute for Energy Research (IREC), Jardins de les Dones de Negre 1, 2, 08930 Sant Adrià de Besòs, Barcelona, Spain
b
Universitat Politecnica de Catalunya – BarcelonaTech (UPC), Jordi Girona, 1-3, 08034 Barcelona, Spain

HIGHLIGHTS GRAPHICAL ABSTRACT

• Regulatory framework on Demand

Aggregators strongly depends on the
country.
• Main barriers for Demand Aggregators
have been identified and analyzed.
• Aenable
regulation scheme is proposed to
Demand Aggregators partici-
pation.
• Small tertiary building aggregation is
still not economic viable.
• Municipal retailers could consider to
extend their operations to Demand
Aggregation.

ARTICLE INFO ABSTRACT

Keywords: European Directives are incentivizing consumers to play an active role in the electricity system and to colla-
Demand Aggregator borate to maintain its stability, which has been historically provided by large generation power plants. However,
Regulatory framework it is not easy for the System Operator to handle the coexistence of consumers and generators in the same markets.
Ancillary Services Under these circumstances, a new actor allows small residential and commercial consumers to participate in
Demand Response
flexibility markets: The Demand Aggregator. However, balancing markets opened to Demand Aggregators still
Tertiary building management
present several barriers that do not allow their practical participation. This study analyzes barriers and enablers
of four European electricity markets and proposes a new market framework that would enhance Demand
Aggregators’ participation. To validate the proposed market and to understand the economic potentials of ag-
gregated small tertiary buildings, a Demand Aggregator is simulated using real building’s consumption data.
Results show that technical requirements to participate in balancing markets such as the minimum bid size, the
symmetricity of the offer and the product resolution strongly affect incomes for Demand Aggregators. However,
neither in the proposed market, the creation of a Demand Aggregator whose business model is focused on small
tertiary buildings does not seem realistic due to low incomes in comparison to the fixed costs necessary to enable
Demand Response, especially if only the air conditioning system is considered.

Corresponding author.
⁎

E-mail addresses: [email protected] (M. Barbero), [email protected] (C. Corchero), [email protected] (L. Canals Casals), [email protected] (L. Igualada),
[email protected] (F.-J. Heredia).

https://doi.org/10.1016/j.apenergy.2020.114707
Received 9 October 2019; Received in revised form 14 February 2020; Accepted 18 February 2020
0306-2619/ © 2020 Elsevier Ltd. All rights reserved.
M. Barbero, et al. Applied Energy 264 (2020) 114707

1. Introduction with the EU policy objectives. Moreover, various countries, such as

Spain, have not yet transposed European Directives having their bal-
The higher penetration of Renewable Energy Sources (RES) and its ancing markets closed to prosumers [17].
inherent stochasticity is changing the way in which electricity is traded Nonetheless, industry, residential and tertiary buildings can become
and managed in the electricity system. The old paradigm in which potentially eligible for participating in DR programs. Although the
production is adapted to consumption is no longer feasible [1]. In this majority of existing aggregators deal with large consumers as industries
new framework, energy production is not directly controllable and [18], literature is focusing on residential and tertiary buildings as well
forecasts are not fully accurate. In addition, an increase in the energy because they represent about 40% of the global energy consumption
consumption implies higher demand peaks and grid congestions, which [19]. Such buildings have several energy consumption systems, such as
could increase grid reinforcement needs [2]. Considering that nowa- Heating, Ventilating and Air Conditioning (HVAC), that can be con-
days grid balancing is vastly done by big power generation plants, trolled to provide demand-side flexibility services [20]. The services
which tend to be less reliable, flexibility offered by end-consumers these prosumers may offer depends on the elements installed in the
through Demand Response (DR) mechanisms is crucial. building. Stationary batteries [21] and electric vehicles (EV) [22]are
There are two DR types: implicit DR where the end-users con- also considered important devices for balancing the grid.
sumption is expected to react to a price signal and explicit DR where This study critically examines current European markets framework
flexibility offered by end-users is traded in energy markets [3]. This regarding DA and DR. Section 2 analyzes the existing business models
study immerses in the latter type to improve the use of energy resources in Europe, differentiating third-party aggregators and aggregators as
and energy processes involved in the grid balancing services. retailers. Then, the section analyzes the market requirements to un-
The latest advances in smart grid and building technologies promise derstand how they could act as a barrier for consumer’s participation
to unlock the participation in DR programs and to transform passive distinguishing among regulatory, technical and economic barriers. To
consumers into active consumers, also called “prosumers”. Prosumers have a clear picture of the situation in Europe, some of the most re-
are able to modify their energy consumption depending on external levant markets opened to DA in Europe [17] (Belgium, Finland, France
signals, i.e. economic or environmental among others. Smart buildings, and UK) are analyzed, highlighting principal enablers and barriers.
having to manage their energy flows of generation from renewable After that, the section examines technologies currently used and the
power sources and consumption, may include solutions to evaluate markets in which aggregators are operating, to understand the con-
their flexible loads. This means that technical barriers are no longer nection between market configuration and flexibility provided by ag-
significant on the automation side. The main challenge is to transform gregators.
these functionalities into products that consumers can trade in elec- Then, with the main objective to propose a possible frequency
tricity markets [4] to reduce their electricity bill while helping the regulation market that enables DR participation, Section 3 proposes an
energy transition toward a 100% renewable energy system. Currently, aggregator owned by the local administration due to the fact that public
although commercial services and residential consumer represent the buildings represent a high share of tertiary buildings in cities and they
major share of electricity consumption, according to the International suppose an important potential source of flexibility. This Section,
Energy Agency [5], their flexibility potential remains untapped. within the framework of the REFER project [23], shows how a tertiary
Demand Aggregator (DA) has emerged as a new market agent ne- building DA would operate in the proposed market based on the in-
cessary to manage demand-side flexibility [6]. Its role is to aggregate formation gathered from the libraries in the Metropolitan Area of
different flexibility providers (loads and/or Distributed Energy Re- Barcelona (AMB, from the Spanish acronym). Flexibility sources from
sources (DER)), allowing them to participate in electricity markets [7]. libraries are HVAC and self-consumption solutions with storage sys-
DA is able to manage its client portfolio directly through contracts tems. The flexibility of the aggregator portfolio is quantified under
(unconditional delivery) or indirectly (conditional delivery) through three different scenarios: the first one considers libraries having air
price incentives [8]. Literature is full of optimal DR strategies for par- conditioning as the only flexibility resource; the second one considers
ticipation in the wholesale market [9]. Moreover, DA could participate that 30% of libraries have installed a self-consumption solution with
in Frequency Containment Reserve (FCR) [10], in Frequency Restora- storage system; the third one includes participation in different markets
tion Reserve (FRR), that is divided in manual (mFRR) and automatic at the same time. Finally, the study performs an economic simulation of
(aFRR), and in the Replacement Reserve (RR) [11]. Centralized power the DA participation in the proposed market to evaluate costs and
plants have historically provided these frequency regulation services. benefits of offering balancing services and analyzes the effect of dif-
For this reason, existing markets are strongly oriented towards gen- ferent technical requirements on the DA business model. Conclusion
erators and, in majority of cases, they do not allow a real participation and final remarks are presented in the last section.
of demand side resources. In a high renewable penetration scenario, it
is fundamental to reassess these markets considering the participation 2. Materials and methods
of small consumers and renewable resources, which have completely
different characteristics compared to large thermal power plants [12]. This section presents an overview on DA in Europe. After a brief
To improve energy processes and use, the Energy Efficiency overview of the DA business models emerging in Europe, possible
Directive 2012/27/EU states that barriers for DR participation have to barriers to spread DA are discussed. Then, this section analyzes prin-
be removed and that DR has to be encouraged, including the partici- cipal European markets that are already open to DA and, finally, this
pation of aggregators [13]. Moreover, the EU winter package “Clean section presents the current situation in Spain and the case study ana-
Energy for All Europeans” previews faster markets, where the energy is lyzed.
traded close to real-time and intraday and balancing markets gain even
more importance. The package suggests to incentivize the use of de- 2.1. Market models for demand response aggregation
mand side flexibility and storage resources, strengthening the role of
the aggregator [14]. Although some countries have already opened the DA business models consist in trading the flexibility of their clients
market to DA, they still maintain several technical requirements to one or more actors through market mechanisms or through bilateral
strongly oriented to classical centralized generation sources, reducing contracts. The DA can participate in frequency regulation services or
potential participation of consumers in the system [15]. A recent study help to solve grid congestions by selling flexibility to the Transmission
[16], which analyzed balancing markets in Austria, Germany and or Distribution System Operator (TSO/DSO). Another option for the DA
Netherlands, found key differences among the countries of study and is to help balancing the Balance Responsible Parties (BRP) and/or re-
presented some examples of how the markets’ design is not yet aligned tailers’ portfolio. The DA could operate outside the conventional chain

2
M. Barbero, et al. Applied Energy 264 (2020) 114707

Fig. 1. Comparison between “Third party aggregator” and “aggregator as retailer” business model.

of energy supply i.e. is neither BRP nor retailer (Third-Party • Inappropriate or incomplete regulation defining roles and respon-
Aggregator) [24], or it can be the same BRP/retailer acting as DA [25], sibilities between market’s participants [29]: TSOs should clearly
as shown in Fig. 1. define the balance responsibility in case of flexibility activation from
In the case of a Third-Party Aggregator, consumers contract the part of a DA. If the TSO does not exclude the activated flexibility
energy provision from different retailers and the DA takes advantage of from the retailer/BRP’s balancing area, a DA can cause unfair pur-
the consumer’s flexibility by selling the aggregated flexibility to dif- chasing and balancing risks to retailers, BRPs and DSOs [6].
ferent actors. In this case, consumers do not buy energy from the DA, its • Number of contracts needed for DR [26]: The need for DA to sign a
role is to trade and manage the consumer’s flexibility. The DA can ei- contract also with the consumer’s BRP/retailer/DSO can be a strong
ther provide frequency regulation services and congestion management barrier as they are potential competitors. In case of incomplete
to the TSO/DSO or balance external BRP/retailer’s portfolio by trading regulation on balancing responsibility’s BRP, retailers and DSOs are
the shifted energy in intraday markets. The main drawback of this not incentivized to allow any DA trade their consumer’s flexibility
business model is that the flexibility activation from part of the DA since DA can create additional costs to them.
could create unbalances in the consumer’s BRP/retailer/DSO portfolio.
Without clear rules about the unbalances created by the DA, the con- Regulatory barriers can forbid or limit the participation of DA in the
sumer’s BRP/retailer/DSO could be penalised unfairly [6] or the DA markets. If the regulatory framework is organized to exclude DAs, ag-
could be indebted for the unbalance created. gregators’ revenues will be null [12].
In the case in which the DA is the same as the retailer, its main Technical barriers are imposed by functional requirements needed
business is to sell energy to its clients. However, in Europe, there are to participate in frequency regulation markets that have been histori-
retailers that act as DA, since they offer special tariffs to consumers that cally defined for generation units and should be updated for allowing
are able to shift part of their consumption when it is necessary. Usually, the participation of DA [15]. Tertiary or residential buildings partici-
they are retailers that own renewable generation assets and are able to pating in DR programs have characteristics completely different from
take advantage from the flexibility of their clients to reduce the un- generators and their major constraint is to assure their occupants’
balances costs by balancing their own portfolio. Complementary, a comfort. It is worth to remember that a market agent can deliver an-
competitive retailer will use DR in order to reduce the risk of being cillary services to the TSO only if it is prequalified [30], demonstrating
exposed to high prices in the spot market [26]. If retailers have enough the capability to respect all technical requirements. For this reason, it is
flexibility, they could also provide frequency regulation services and very important that prequalification is made at the DA portfolio level. If
congestion management to TSO/DSO or balance external BRP’s port- prequalification is made at an asset level, each consumer has to be able
folio. The main drawback of a retailer as DA is that it can raise some to respect all the market’s technical requirements on their own [29].
conflicts of interest. The requirements are:
In our case of study, an innovative public aggregator as retailer is
proposed. The AMB has already started one of the first municipal re- • Minimum bid size: indicates the MW necessary to participate in the
tailers, called “Barcelona Energia” [27], and could expand its business market. If this requirement is lower, than the DA needs fewer cus-
model with DA. In this case, the DA, a part from being a retailer, could tomers to participate [31].
also be the owner of the buildings and offer frequency reserves to the • Maximum number of activations: indicates the maximum number of
TSO, taking advantage of its building’s flexibility. The advantage of the time that a flexibility resource can be activated during a certain
business model proposed is that DA benefits are not shared with the period. DA consumers have restrictions about the maximum number
final users, allowing keeping more profitable a business with low of activations during a period to maintain comfort constraints.
margins. • Symmetricity of the offer: flexibility can be in two directions, up-
ward or downward regulation. If the offer needs to be symmetric,
2.2. Frequency markets and barriers the number of consumers that can participate in DR is lower, given
that some consumers can offer flexibility just in one direction [31].
Different type of barriers have been analyzed for DR in frequency • Notification time: indicates the maximum reaction time of the
regulation markets [28]. This study follows the idea presented in [12], flexibility source. Short notification time can give raise to problems
which grouped possible barriers for DA to entry in frequency regulation due to the communication delay between the DA and the consumers’
markets in three types: regulatory, technical and economic barriers. reaction time, apart from increasing automation costs [31].
Regulatory barriers refer to all those barriers that can appear due to • Duration of delivery: Shorter the maximum duration of the flex-
the market regulated and not-regulated framework, that are: ibility activation, more consumers are able to participate in the
service, since most of residential and tertiary consumers can activate
• Restriction on demand aggregation: Although Demand aggregation flexibility as maximum during 1 or 2 h [32].
is allowed, there can be still restrictions on the type, the size or the • Product resolution: Indicates the minimum time during which a unit
voltage connection of the load [12]. has to offer its flexibility. If it is very long, e.g. one day, it can limit

3
M. Barbero, et al. Applied Energy 264 (2020) 114707

DR participation, since different consumers could offer their flex- 2.3. Market analysis
ibility just during some hours a day [32].
• Tender period: Indicates how often the market opens. If there is not Taking into account the previous market description and barriers,
a daily auction it could be difficult for the DA to predict the flex- this section presents an overview on frequency markets opened to DA in
ibility of its clients [32]. Belgium, Finland, France and UK. Prices come from the ENTSOE’s
transparency platform [35].
Technical requirements can limit the available reserve of DAs in the
market and can limit the type of consumers that can participate in the 2.3.1. Belgium
DA portfolio, reducing their profitability. Belgium increased DR programs after important capacity shortages
Economic barriers occur when the DA business model is not viable due to technical issues of some nuclear power plants in the country in
due to costs exceeding benefits from participation of DR in balancing the last years. Moreover, the planned closure of some conventional
markets [33]. These barriers are: power plants and nuclear power plants and the increase of renewable
capacity [36] makes DR a vital source for the system.
• Low prices in frequency regulation markets.
• High technical costs: smart meter installation, communication and • Principal enablers:
control technologies, automation, etc… can reach high costs if very o Third-party aggregators can participate in the market.
high performances are demanded to participate in the markets [34]. o Offers do not need to be symmetrical in FRR and RR.
• High penalization costs: Market costs such as penalization for not o The minimum time between two successive activations is 8 h in
dispatching the committed energy should be reduced to incentivize the mFRR market.
demand side participation [28]. o Prequalification takes place at pool level.
• Subsidies to peak power plants: they can create an unfair competi- o For FCR and FRR penalties are proportional to the payments, with
tion; peak power plants are the direct competitors in provide bal- a multiplication factor of 1.3.
ancing services to the grid. The absence of direct incentives to DR • Principal barriers:
technologies could decrease revenues for DA. o DSOs can block the consumer participation in DR programs
without taking responsibility for the costs incurred by the con-
Economic barriers due to the market design affect the way in which sumer, DA and TSO.
the same reserve will be remunerated. A good market design should o Contracts are made on yearly basis for RR.
assure a fair remuneration to DA and give incentives to provide services
to the network. Table 1 illustrates technical requirements described by Elia, the
The barriers presented have different links among them, as re- Belgian TSO [37].
presented in Fig. 2, which should be taken into account to improve the
balancing market design. At first, regulatory barriers should be avoided 2.3.2. Finland
to allow DA participation. Then, technical requirements need to assure Finland has the necessity to add flexibility in its grid as currently the
participation to the largest pool of flexible loads to maximize their major part of the Finnish reserves are bought from its neighboring
availability. Finally, a good market design is necessary to allow a suf- countries such as Estonia, Sweden, Norway and Russia [38]. DA can
ficient remuneration to DA and reduce their financial risks. help Finland to be more independent from these countries.

• Principal enablers:
o Unbalances created by the DA in a BRP area does not increase

Fig. 2. Links between barriers for DA in balancing markets in hierarchical order of importance.

4
M. Barbero, et al. Applied Energy 264 (2020) 114707

Table 1
Summary of balancing markets open to Demand Aggregators in Belgium.

1
Dark cells represent that the requirement should be improved to increase DR participation.

costs for the BRP, as the TSO corrects the BRP curve after the DA 2.3.4. UK
flexibility activation. Although UK was one of the first countries to incorporate DR so-
o Prequalification takes place at portfolio level. lutions in Europe, the market is yet immature and the capacity of DR is
o Smart meters are widely used. decreasing each year, risking to disappear in the future [6].
o The minimum bid size is 0.1 MW for FCR-N.
o Product resolution is 1 h for all services. • Principal enablers:
• Principal barriers: o DA can access consumers directly without the permission of the
o DA needs the agreement of the consumer’s retailer/BRP. BRP/retailer.
o Aggregating sources from different BRP’s areas is only allowed in o Prequalification takes place at pool level.
FCR market. o The maximum RR (STOR for the national TSO) activations per day
o The minimum bid size is 5 MW for FRR and 10 MW for RR ser- is agreed with the TSO.
vices. o The system used for counting grid charges to consumers can help
DA business model (TRIAD system).
Table 2 illustrates technical requirements described by Fingrid, the o A part from utilization and capacity payment, balancing service
Finnish TSO [39]. providers get also the nomination payment, which consist in a
holding fee for each hour (£/h) used within nominated windows.
o In 2018, UKPN presented their Flexibility Roadmap, an ambitious
2.3.3. France plan to develop market-based solutions to procure flexibility for
France is possibly the European country with the longest tradition in its network where DA can participate[42].
DA, along with the UK. The massive presence of nuclear power plants • Principal barriers:
and the wish to increase renewable generation resulted in a great in- o Tender period can be a barrier in all markets.
terest of the country in DR programs. However, prices of balancing o The minimum bid size in the aFRR market is 25 MW.
markets are dropping in last years, making more difficult the business o Demand Turn Up service did not take place in 2019.
for DA.
Table 4 illustrates technical requirements described by National-
• Principal enablers: Grid, the UK TSO [43].
o DA can access consumers directly without the permission of the Table 5 summarizes some of the current business models in the
BRP/retailer. countries analyzed. France, UK and Finland are the only countries were
o The “appel d’offres Effacement” (RR) is appositively thought for some residential consumers are aggregated in Europe. In general, the
consumers [40]. great majority of DA works with industrial or large energy consumers,
o Prequalification takes place at pool level. Voltalis is the only DA that works exclusively with households. In UK,
o The duration of delivery is well suited for consumers for all ser- where consumers are charged depending on their consumption during
vices. the three peaks power of the country during the year, all DA analyzed
• Principal barriers: try to reduce grid charges. However, strict requirements for FRR in UK
o Aggregation of DR and generation in the same bid is not allowed. block the entrance of DA in that market. In Finland, all DA analyzed
o Generators are obligated to deliver a-FRR services, however they participate in FCR markets because they are well suited for consumers.
can subcontract DR services through secondary markets. All DA having industries in their portfolio participate in RR markets
o Participation in aFRR market is limited to that consumers con- because they are the best suited for large energy consumers.
nected at the TSO level.
o The minimum bid size for mFRR services is 10 MW. 2.4. Current framework in Spain
o FRR and RR are tendered on yearly basis.
The case study of this work is located in Spain; therefore, a specific
Table 3 illustrates technical requirements described by RTE, the analysis of the Spanish framework is required. In Spain the day-ahead,
French TSO [41]. intraday and future electricity markets are managed by OMIE [44],

5
M. Barbero, et al. Applied Energy 264 (2020) 114707

Table 2
Summary of balancing markets open to Demand Aggregators in Finland.

while ancillary services are managed by the national SO, REE [45]. the national a-FRR, while the other flexibility mechanisms are an ob-
Nowadays, DR is allowed just to large energy consumers (5 MW) ligation for all generation plants. Power plants are paid for their
through the interruptibility services [46]. This program has not been availability during peak hours and the cost is a bit more than 10,000
activated for several years raising questions whether it is a genuine €/MW per year, representing about the 5% of the electric tariff in Spain.
interruptible load program or a form of subsidy to the national industry This is not an efficient system from an economic point of view; in ad-
[6]. Furthermore, the only market mechanism for trading flexibility is dition, it does not accomplish the guidelines of the Energy Efficiency

Table 3
Summary of balancing markets open to Demand Aggregators in France.

6
M. Barbero, et al. Applied Energy 264 (2020) 114707

Table 4
Summary of balancing markets open to Demand Aggregators in UK.

Directive 2012/27/EU and of the Winter Package. For these reasons, and demand. The addition of DR sources in the system is a necessity for
the analysis performed through European markets already opened to the country to hold a transition toward a 100% renewable electricity
DA is useful to formulate a proposal that can be used for the Spanish system.
one.
In Spain, 49% of power installed in the country comes from re- 2.5. Case study
newable sources, being the 30% from solar and wind [47]. The high
share of stochastic sources in the generation mix means high needs for The case study considers a DA of tertiary buildings based on the
flexibility in the grid. Today, to assure that the generation can cover all characteristics of the 61 libraries of the AMB, in order to simulate the
the demand, there are 108 MW installed of power plants against the behavior of a public DA in the market framework proposed for Spain.
historical maximum demand registered of 45 MW. DR could be a AMB has recently started one of the first municipal retailers in Spain
cheaper and more environmental-friendly solution to reduce payments [27]. The possibility to take advantage of the flexibility of its own
to peak power plants and decommission the older ones. Moreover, wind buildings as libraries, schools and offices is a great opportunity for the
curtailment grew exponentially from 2008 to 2013 with an economic AMB. DA could help to accomplish local climate and energy targets set
impact of around 85 M€ [48] due to a mismatch between generation by the Covenant of Mayors [49]and at the same time innovate the

Table 5
Analysis of the main European Demand Aggregator business model.
Aggregator FCR FRR RR Wholesale/intraday Reduction of grid Portfolio Client target Act as
market charges balancing retailer

Belgium Restore [56] X X X X Industries, tertiary buildings

Yuso [57] X X Renewables, batteries, industries X
Finland Seam [58] X X X Large energy consumers
Fortum [59] X X Households batteries, EVs, X
renewables
France Smart Grid energy X X X Industries, generators
[60]
Energy Pool [61] X X X X X Industries, DER
Voltalis [62] X X X Households
UK Open Energi [63] X X X X X Industries, generators, batteries.
Kiwi Power (Kiwi X X Industries, tertiary buildings,
[64] batteries and CHP
Flexitricity [65] X X X X X CHP, consumers, batteries, back-up
generators, renewables

7
M. Barbero, et al. Applied Energy 264 (2020) 114707

retailer business models, possibly increasing revenues through a public HVAC consumption (grey line) is reduced with respect to the baseline
DA/retailer. (yellow line), calculated as the interpolation of the consumption be-
Data are gathered from the library situated in Montgat (Spain) from tween the start and the end of the flexibility activation. The estimated
October 2017 to October 2018 [50], while the other libraries are si- up flexibility (orange line) represents the expected minimum con-
mulated depending on their real characteristics. Flexibility sources in sumption that the library could have, which is very close to the reached
the Montgat library are the HVAC and a self-consumption solution value.
composed by a PV panel and a second life electric vehicle battery. The The storage system installed in the library is a second life EV
accuracy in electric measurements responds to the characteristics de- battery, with capacity of 18.4 kWh and the power is limited to 10 kW
termined by the IEC (International Electrotechnical Commission) by the converter. The strategy used in the building is to charge the
Standard 62053-11 [51] for Class 0.5, which is 0.5% under full load battery during the night, when the energy is cheaper, to use that en-
conditions. ergy during the day, from 11 h on. Therefore, the battery can offer
The HVAC installed in the library is a Neptuno 125 by Ferroli, with downward regulation during the day and upward regulation during
a power peak of 39 kW and able to produce 126 kW of heat and 116 kW the night.
of cold. The HVAC behavior is simulated with a multi zone building that The HVAC system and the battery, due to their physical character-
calculates the hourly consumption of the library using external tem- istics are suitable for participating in FRR services, while they could not
perature, solar irradiation and the set-point temperature. Knowing that participate in the RR market due to the large duration of delivery. The
the range of comfortable temperatures is the set-point temperature ± battery can also participate in the FCR market by continuously injecting
1 °C, it is possible to calculate the flexibility of the library by changing or consuming energy from the grid whenever the frequency is lowing or
the set-point temperature in the allowed range. The simulation uses increasing respectively [54].
Type 56 [52] of TRNSYS® and, although it is a simplified model, results To estimate the flexibility of each one of the 61 libraries of the AMB
are coherent with other studies [53]. To validate the model, a flexibility the ratio between the power contracted by the Montgat library and each
activation from 15:00 to 15:45 was simulated during the 15/07/2019 in one of the other libraries is used, assuming that the same proportion
Montgat library. From the simplified model, in July the library can corresponds to the flexibility available at libraries. Then, using the
reduce its HVAC consumption of 10.8 kWh when it is open during one timetable of each library, the DA portfolio flexibility was calculated.
hour. Fig. 3 shows the behavior of the library during the activation: the Notice that the study takes into account that the flexibility is available
since one hour before the opening of the library, as the building can be
pre-heated or pre-cooled during that period.
The case study considers three different scenarios:

1. Libraries as they are. All buildings have HVAC but just one library
has a battery. In this scenario, all the flexibility is traded in the FRR
market.
2. 30% of the libraries have electric batteries. As shown in Table 5
storage systems are often expected in the DA business models.
Again, all the flexibility is traded in the FRR market.
3. Participation in FCR and FRR markets. Based on the specifications of
Scenario 2, this 3rd Scenario considers that 20% of the flexibility
offered by batteries is traded in the FCR market, while the remaining
80% is traded in the FRR market.

Fig. 3. Simulation of a flexibility activation in July in the library. Fig. 4 represents upward and downward flexibility for a typical day
in January for Scenario 1 and 2.

Fig. 4. Libraries’ flexibility in Scenario 1 and in Scenario 2 during 22th January.

8
M. Barbero, et al. Applied Energy 264 (2020) 114707

To calculate benefits from DR participation, the TSO calls to the

DA are simulated using Matlab®, following the market characteristics
described in Section 3.1. In detail, the number of calls per each day
and direction comes from the round of a Normal distribution with an
average equal to the average number of activations and a standard
deviation equal to 0.7. Similarly, the duration of the activation varies
randomly between the minimum and the maximum duration of de-
livery. The prices for utilization of the upward and downward reg-
ulation are those from the current FRR market in Spain during 2019
[45]. Regarding the capacity payment for FRR, the study assumed a
price of 3.3 €/MW, according to the average price in the Finland
market during 2019. Regarding capacity payments in the FCR market,
it has been assumed the average price of the European symmetric FCR
200 mHz during 2019, that is 8.6 €/MW/h, where France and Bel-
gium are participating. Finally, the hours activated are the hours with
the highest price in the current market for two reasons: 1) The hours Fig. 5. Comparison among countries and the proposed market for FCR balan-
with the highest prices are the hours in which the grid is more cing services.
stressed, so it is when a service that is activated just few times a day
would be used; 2) This allows to keep into account eventual spike
Regarding mFRR, the notification time recommended is 15 min, in
prices in the market.
line with all markets analyzed. Taking the Belgian example, it is pro-
To calculate the DA’s benefits, it is assumed that the flexibility
posed a minimum bid size of 1 MW, because higher minimum bid sizes
offered to the markets is the 80% of the actual flexibility
could be difficult to reach for a tertiary building DA. In order to allow
calculated, to ensure that the committed capacity is delivered by
tertiary building DR, the average number of activations per day re-
the DA even when some individual consumers may not be able to
commended is two, as in France. 30 min product resolution is sug-
perform in order to avoid penalizations [17]. In this study the DA is
gested, as a longer product resolution would reduce the number of
considered as a price taker that uses hourly marginal prices and,
clients able to deliver the service using HVAC or batteries. In all
thus, the formulation of an optimal bidding strategy is out of the
countries analyzed, mFRR is not a symmetrical service and the duration
scope.
of the service is between 15 min and 2 h. Duration of delivery between
15 min and 1 h are well suited for tertiary buildings. Regarding the
3. Discussion and results
tender period, in all countries apart from Finland, where it is contracted
until 45 min before the hour of use, FRR is tendered monthly or yearly.
According to the qualitative analysis of best practices, enablers and
The Finnish case demonstrates that shorter tender time is possible, a
barriers from existing markets in Europe, this section proposes a na-
daily tender can be a good trade-off for allowing participation of DR in
tional frequency energy market followed by an economic simulation of
the market, taking into account that it could be very complicated for the
the functionality of a tertiary buildings’ DA in the market.
DA to predict the flexibility one month or one week ahead. Marginal
price for capacity payments and bid price for utilization payments
3.1. Proposed frequency market
should reflect costs and ensure revenues to all market participants, as
already happens in France and Finland. Fig. 6 represents qualitatively
In order to allow participation of small consumers in frequency
the FRR market proposed in respect to the four markets analyzed.
markets, it would be desirable that the prequalification takes place at
pool level, as in Finland, UK and France. Otherwise, just large energy
consumers will be able to be prequalified and DA would not increase
the number of consumers that could participate in frequency markets.
Taking Belgium, France and UK as example, it would be better if DA
does not sign any contract with BRP/retailers and DSOs but directly
with prosumers. In addition, the TSO should automatically adjust the
BRP/retailer’s curve when flexibility is activated from part of the DA, as
it is done in Finland, to avoid to increase BRP/retailer unbalances costs
due to the DA action.
FCR, due to the nature of the service, is the most similar market
among countries. It is a very rapid regulation, for this reason the
notification time is between 2 and 15 s and it is activated con-
tinuously. The Finnish case is a great example of how, at least in FCR
markets, the minimum bid size can be 0.1 MW. As in Finland, a pro-
duct resolution of 1 h is proposed with daily auctions. Capacity pay-
ments are necessary for this type of service and should be higher than
in other markets, as it is a more sophisticated service. Fig. 5 represents Fig. 6. Comparison among countries and the proposed market for FRR balan-
qualitatively the FCR market proposed in respect to the four markets cing services.
analyzed.

9
M. Barbero, et al. Applied Energy 264 (2020) 114707

Regarding RR, the notification time varies significantly among

Capacity payment

Marginal bid price

countries; it goes from 15 min to 8 h. DR sources usually can react

Marginal price
Marginal price
relatively fast, a minimum notification time of 2 h as in France or less
can be enough for DA. The minimum bid size should be fixed to 1 MW
as in Belgium and UK to boost DA participation. UK is a great example
of how the maximum number of activations can be agreed with the DR
source. The duration of delivery should be fixed to 2 h, as in France. In
France, Finland and UK the product resolution of 1 h allows catching

Utilization payment
DR potentials, the same product resolution is proposed here. In all
markets analyzed these reserves are contracted yearly or seasonally. A
daily tender period would facilitate flexibility forecast for the next

Bid price
Bid price
period, and indeed, it would enhance consumer’s participation. Finally,
in the market proposed there should be both capacity and utilization

0
payments as in UK, France and Belgium. Fig. 7 represents qualitatively
the RR market proposed in respect to the four markets analyzed.

Tender period

Daily
Daily
Daily
Duration of delivery

15 min to 1 h
No stop

2h
Symm

YES
NO
NO
Product resolution

Fig. 7. Comparison among countries and the proposed market for RR balancing
services.

Table 6 shows the summary of the technical requirements proposed

30 min
1 Hour

for the different services.

1h

3.2. DA profits
Indicated by the service provider
Max. number of activations

This section shows the possible profits that the DA would have in
the proposed FCR and FRR markets for our case study and the total
Continuous activation

energy shifted due to the balancing services offered. In addition, the

2/day on average

study evaluates both the number of calls accepted by the DA and the
minimum number of libraries needed to assure profits to the DA’s cli-
ents. The number of calls accepted represents how many times the DA
has flexibility available and it is activated by the TSO. Notice that if the
DA has no flexibility available it will not do any bid to the market and
the TSO will use other flexibility providers to restore the frequency in
Proposed balancing market open to Demand Aggregators.

15 s 50% 30 s 100%

the grid. The number of libraries needed are calculated to reach the
95% of the profits, taking into account just the hours in which the
flexibility offered is higher than the minimum bid size. From the
Not. time

Matlab® simulation, the number of downward and upward FRR acti-

15 min

vations was 668 and 665 times respectively, with an average duration
2h

of the activation of about 38 min.

In the first scenario, libraries can respond to 531 calls out of 1333
Min. bid size [MW]

from the TSO to the DA. It is assumed that during these hours the DA
can activate other clients and, if not, it would not have offered flex-
ibility into the market. Revenues from utilization and capacity are 5323
€ and 5835 € respectively, which make a total amount of 11,158 €. The
total energy shifted to provide balancing services is respectively 65
0.1

MWh and 44 MWh for downward and upward regulation.

1
1

With these market conditions, the DA would need at least 596 li-
braries to assure its participation in the market. With less libraries, the
Market
Table 6

mFRR

DA would not reach the minimum bid size required.

FCR

RR

In Scenario 2, libraries can respond to 985 calls thanks to the

10
M. Barbero, et al. Applied Energy 264 (2020) 114707

Table 7 Table 8
Results by Scenario in the proposed balancing market. Revenues and number of libraries needed under tuned technical requirements.
Scenario 1 Scenario 2 Scenario 3 Tuned parameter Revenues per Number of
library [€] libraries needed
Number of calls accepted 532 985 985
Utilization Payments [€] 5323 13,699 12,025 Base case 457 168
Capacity Payments [€] 5835 14,230 16,925 Minimum bid size = 5 MW 457 841
Total energy shifted for FRR service 109 276 243 Minimum bid size = 0.1 MW 457 17
[MWh] Average number of activations = 4 614 171
Number of buildings needed to reach 596 168 210 Average number of activations = 1 351 168
the minimum bid size Symmetric offer 160 671
Product resolution = 4 h 316 366
Product resolution = 24 h 0 NA

telecommunication and monitoring, according to [33]. This means that

total benefit per library per year are respectively 48, 323 and 341 €.
Considering that the average annual bill of the libraries analyzed is
about 30,000 €, possible revenues represent a very low percentage of
the total bill. Fig. 8 shows the results.
Notice that in this study costs and benefits are considered as an
overall and that taxes are not counted. If from these results taxes and
the share of the profits among all clients should be subtracted from the
DA profits, the business model would not probably be viable. However,
hardware costs for DR could be amortized by the reduction in the en-
ergy bill due to the monitoring and the consumption optimization of the
building. Including batteries in the study substantially increases the
brut profit of the DA. However, although the investment costs for bat-
teries and inverters are not considered due to the inherent complexity
of battery ageing according to its working conditions [55] and, thus, the
amortization period, this profit is expected not to be enough to guar-
antee the payment of a battery replacement, as shown in a recent study
from the same authors that already analyzed costs and benefits of using
second life batteries for DR in balancing markets here [54].
In order to better understand the impacts of the technical require-
ments on the DA business model, Table 8 shows benefits and total
Fig. 8. DA costs and profits by Scenario in the proposed balancing market. number of buildings needed for Scenario 2 under tuned market condi-
tions. Values for the base case come from the market conditions re-
presented in Table 6.
contribution of batteries, considerably more than in Scenario 1. This is The minimum bid size of the offer has no impact on the revenues
because batteries are available also when the library is closed, while the generated. However, it has a strong influence on the number of libraries
HVAC is available just when the library is open. In this case total rev- needed by the DA to assure that revenues. There is almost a linear
enues would rise to 27,929 €, of which 13,699 € comes from utilization correlation between the number of libraries needed and the minimum
and 14,230 € from capacity. The total energy shifted increases when bid size of the offer. This means that if the minimum bid size is higher,
batteries are considered. In this scenario 162 MWh and 114 MWh for the DA needs a larger portfolio to participate in the frequency regula-
downward and upward regulation respectively where used. This result tion market.
shows how batteries increase substantially the flexibility of buildings, The average number of activations mostly influence revenues. While
reducing also the number of buildings needed by the DA to reach the there is a fixed part (capacity payments) that does not depend on the
minimum bid size. In this case, at least 168 tertiary buildings would be total number of activations, the variable part (utilization payments)
necessary to reach the minimum bid size, assuring the 95% of the does. By increasing the number of activations, revenues per each library
benefits described. grow. However, the comfort of the building’s occupants is affected more
In Scenario 3, libraries can also respond to 985 calls. In this third times per day and buildings’ managers would be less willing to take part
scenario, revenues come from two markets, resulting in 12,025 € from of the DA portfolio.
utilization and 12,550 € from capacity payments from FRR services, It is found that the symmetricity of the offer is a very important
4375 € from capacity payment from the FCR market, generating a total parameter in the DA business model. By putting this condition, the DA
income of 28,950 €. In Scenario 3 DA provides 143 MWh and 100 MWh is forced to offer ever the lower available flexibility between the up-
for downward and upward regulation respectively for FRR services. As ward and the downward, reducing the total flexibility offered to the
buildings participate in two markets at the same time, the flexibility system. Revenues are dramatically reduced (divided by three) and the
needed to reach the minimum bid size for both services at the same number of buildings needed to assure the participation in the frequency
time is higher and, consequently, at least 210 tertiary buildings would regulation market is multiplied by four.
be necessary. Note that as FCR services pay just for capacity, energy The product resolution affects both revenues and the number of li-
revenues are lower than in Scenario 2. However, capacity payments in braries needed. Increasing the product resolution to 4 h forces the DA to
FCR are higher than in all other cases, making Scenario 3 the most offer the minimum upward and downward flexibility available in the
suitable one, as benefits are 1020 € higher than in Scenario 2. 4 h’ block. In this case, revenues are reduced to 316 € and the number
Table 7 resume the main results of each Scenario. of libraries needed is more than doubled. In the case of a product re-
Average yearly revenue by scenario are 182, 457 and 475 € for each solution of 24 h, the business model would not be possible, because, as
library. The estimated investment cost for DA client is about 134 € per represented in Fig. 4, the upward or the downward flexibility are equal
consumer and per year to cover costs for automation,

11
M. Barbero, et al. Applied Energy 264 (2020) 114707

to 0 in many occasions. Declaration of Competing Interest

From the analysis performed, even though regulatory and technical
barriers to participate through small aggregated tertiary buildings in The authors declare that they have no known competing financial
frequency regulation markets are avoided using the market proposed, it interests or personal relationships that could have appeared to influ-
can be difficult for DA to find a valuable business model by aggregating ence the work reported in this paper.
small tertiary buildings. Economic barriers are yet strong, as prices are
too low to face actual costs in communication and automation needed Acknowledgement
to participate in these markets. However, it is highlighted the im-
portance to facilitate the entrance of small tertiary buildings DA in the The authors gratefully acknowledge the project REFER (COMRDI15-
frequency regulation market, as they can be an important resource for 1-0036), funded by ACCIÓ and the European Regional Development
the grid. Moreover, the impact of the technical requirements on the DA Fund (FEDER) under the RIS3CAT Energy Community. This research
business model is quantified. has also been financially supported by the research and innovation
programme Horizon 2020 of the European Union under the grant
agreement nr. 731211 SABINA. C. Corchero work is supported by the
4. Conclusions
grant IJCI-2015-26650 (MICINN). Prof. F.-Javier Heredia was sup-
ported by the grant RTI2018-097580-B-I00 of the Ministry of Science,
This study identifies the main barriers for small tertiary buildings’
Innovation and Universities of Spain.
Demand Aggregators to participate in frequency regulation services. In
All researchers have been partially supported by the Generalitat de
particular, regulatory, technical and economic barriers are identified in
Catalunya, Spain (2017 SGR 1219). The authors also acknowledge the
hierarchical order. These barriers are reinforced when prequalification
suggestions from the Energy Pool team.
is made at the asset level and when there is a lack of regulation for
determining the effect of Demand Aggregator on the Balance
Appendix A. Supplementary material
Responsible Party or retailer’s portfolio balancing. With the aim to
avoid all these barriers, the study proposes a possible market frame-
Supplementary data to this article can be found online at https://
work.
doi.org/10.1016/j.apenergy.2020.114707.
Economic results from the simulation of a Demand Aggregator ag-
gregating public libraries in the proposed market shows that possible
References
revenues are relatively low, especially when only air conditioning
system is used for Demand Response services. Results highlight that
[1] Khan ASM, Verzijlbergh RA, Sakinci OC, De Vries LJ. How do demand response and
batteries can substantially increase revenues for Demand Aggregators electrical energy storage affect (the need for) a capacity market? Appl Energy
and reduce the number of clients necessary to reach the minimum bid 2018:39–62. https://doi.org/10.1016/j.apenergy.2018.01.057.
size required although the amortization costs of these batteries might [2] Spiliotis K, Ramos Gutierrez AI, Belmans R. Demand flexibility versus physical
network expansions in distribution grids. Appl Energy 2016;182:613–24. https://
significantly reduce this economic advantage. From the simulation also doi.org/10.1016/j.apenergy.2016.08.145.
emerges that Demand Aggregators can increase their revenues offering [3] SEDC (2016) ‘Explicit and implicit demand-side flexibility complementary ap-
flexibility in different markets at the same time. In addition, the study proaches for an efficient energy system explicit and implicit demand-side flexibility:
complementary approaches for an efficient energy system’, (September). Available
analyzes the effect of technical requirements in the Demand at: http://www.smarten.eu/position-papers-reports/.
Aggregator’s business model. It is found that the symmetricity of the [4] Iria J, Soares F, Matos M. Optimal supply and demand bidding strategy for an ag-
offer and the product resolution are the most important requirements to gregator of small prosumers. Appl Energy 2018;213(September):658–69. https://
doi.org/10.1016/j.apenergy.2017.09.002.
evaluate possible revenues and the number of buildings needed by the
[5] Energy in buildings and communities programme (2016) ‘EBC Annex 67 Energy
Demand Aggregator. Flexible Buildings, http://www.iea-ebc.org/projects/ongoing-projects/ebc-annex-
If revenues should be divided among clients and subtracting taxes, 67/’.
[6] Bertoldi P, Zancanella P, Boza-Kiss B. Demand Response status in EU Member
the creation of a business model for small tertiary buildings Demand
States, Europa. eu: Brussels, Belgium; 2016. doi: 10.2790/962868.
Aggregator does not seem possible and a great flexibility potential [7] Behrangrad M. A review of demand side management business models in the
would be blocked. In order to reduce costs, municipalities that already electricity market. Renew Sustain Energy Rev 2015;47:270–83. https://doi.org/10.
participate in electricity markets through a municipal retailer, as it is 1016/j.rser.2015.03.033.
[8] Richter LL, Pollitt MG. Which smart electricity service contracts will consumers
the case in Barcelona with “Barcelona Energia”, could consider to ex- accept? The demand for compensation in a platform market. The Authors Energy
tend their operations to Demand Aggregation. For this reason, this Econ 2018;72:436–50. https://doi.org/10.1016/j.eneco.2018.04.004.
study proposes an innovative public Demand Aggregator, which is also [9] Abapour S, Mohammadi-Ivatloo B, Tarafdar Hagh M. Robust bidding strategy for
demand response aggregators in electricity market based on game theory. J Cleaner
the owner of the flexible buildings. In Barcelona, the municipal retailer Prod 2020;243:118393. https://doi.org/10.1016/j.jclepro.2019.118393.
can act as Demand Aggregator using the flexibility of its buildings to [10] Yuen C, Oudalov A, Timbus A. The provision of frequency control reserves from
balance its own portfolio and selling services to the grid. There is a big multiple microgrids. IEEE Trans Ind Electron 2011;58(1):173–83. https://doi.org/
10.1109/TIE.2010.2041139.
potential considering the great number of public buildings that public [11] Heleno M, Matos MA, Lopes JAP. A bottom-up approach to leverage the partici-
institutions manage. Probably it is the only way to take advantage of pation of residential aggregators in reserve services markets. Electr Power Syst Res
the flexibility of these small tertiary buildings, also considering that 2016;136:425–33. https://doi.org/10.1016/j.epsr.2016.03.025.
[12] Borne O, Korte K, Perez Y, Petit M, Purkus A. Barriers to entry in frequency-reg-
public administration does not need to respond just to market forces, as
ulation services markets: Review of the status quo and options for improvements.
they also need to respect environmental targets and be an example for Elsevier Ltd, 81(July 2016). Renew Sustain Energy Rev 2018;1:605–14. https://doi.
citizens. org/10.1016/j.rser.2017.08.052.
[13] European Parliament (2012) ‘Directive 2012/27/EU of the European Parliament
and of the Council of 25 October 2012 on energy efficiency’, Official Journal of the
CRediT authorship contribution statement European Union Directive, (October), pp. 1–56. doi: 10.3000/19770677.L_2012.
315.eng.
[14] European Commission. Clean energy for all Europeans; 2016. Available at: https://
Mattia Barbero: Methodology, Software, Investigation, Writing - ec.europa.eu/energy/en/topics/energy-strategy-and-energy-union/clean-energy-
original draft. Cristina Corchero: Conceptualization, Validation, all-europeans.
[15] Sweco, Ecofys, Tractbel Engineering and PWC. Study on the effective integration of
Writing - review & editing, Supervision, Funding acquisition. Lluc demand energy recourses for providing flexibility to the electricity system. Final
Canals Casals: Writing - review & editing, Visualization, Supervision. report to The European Commission; 2015, (April), p. 179.
Lucia Igualada: Conceptualization, Validation, Visualization, Project [16] Poplavskaya K, de Vries L. Distributed energy resources and the organized balan-
cing market: A symbiosis yet? Case of three European balancing markets. Energy
administration. F.-Javier Heredia: Supervision.

12
M. Barbero, et al. Applied Energy 264 (2020) 114707

Policy 2019:264–76. https://doi.org/10.1016/j.enpol.2018.11.009. [39] Fingrid. Reserves and balancing power; 2019. Available at: https://www.fingrid.fi/
[17] SEDC. Explicit demand response in europe mapping the markets 2017; 201. p. en/electricity-market/reserves_and_balancing/ [accessed: 24 December 2019].
1–223. Available at: http://www.smarten.eu/wp-content/uploads/2017/04/SEDC- [40] EnergyPool. Lancement de l’appel d’offresis Effacement; 2018. Available at: https://
Explicit-Demand-Response-in-Europe-Mapping-the-Markets-2017.pdf. www.energy-pool.eu/fr/lancement-de-lappel-doffres-effacement-2018/ [accessed:
[18] Shoreh MH, Siano P, Shafie-khah M, Loia V, Catalão JPS. A survey of industrial 14 June 2018].
applications of demand response. Electric Power Syst Res 2016;141:31–49. https:// [41] RTE ancillary services; 2019.
doi.org/10.1016/j.epsr.2016.07.008. [42] Venegas FG, Petit M. Can DERs fully participate in emerging local flexibility tenders
[19] Lindberg KB. Impact of zero energy buildings on the power system - A study of load ? Eem 2019. IEEE; 2019. p. 1–5.
profiles, flexibility and system investments; 2017. Available at: http://hdl.handle. [43] Nationa grid. Balancing services; 2019.
net/11250/2450566. [44] OMIE; 2020. Available at: http://www.omel.es/inicio [accessed: 10 February
[20] Ding Y, Cui W, Zhang S, Hui H, Qiu Y, Song Y. Multi-state operating reserve model 2020].
of aggregate thermostatically-controlled-loads for power system short-term relia- [45] Red Electrica de España; 2020. Available at: http://www.ree.es/es/ [accessed: 5
bility evaluation. Appl Energy 2019:46–58. https://doi.org/10.1016/j.apenergy. February 2020].
2019.02.018. [46] Ministry of Energy. Boletón Oficial del Estado. Boletín Oficial del Estado; 2018. p.
[21] Malhotra A, Battke B, Beuse M, Stephan A, Schmidt T. Use cases for stationary 61561–7. Available at: https://transparencia.gob.es/servicios-buscador/contenido/
battery technologies: A review of the literature and existing projects. Renew Sustain normativaordenministerial.htm?id=NORMAT_E0492160119036&lang=es&
Energy Rev 2016;56:705–21. https://doi.org/10.1016/j.rser.2015.11.085. fcAct=2019-03-15T08:25:41.875Z.
[22] Peng C, Zou J, Lian L, Li L. An optimal dispatching strategy for V2G aggregator [47] entsoe.eu. Installed Capacity per Production Type; 2018. Available at: https://
participating in supplementary frequency regulation considering EV driving de- transparency.entsoe.eu/generation/r2/installedGenerationCapacityAggregation/
mand and aggregator’s benefits. Appl Energy 2017;190:591–9. https://doi.org/10. show?name=&defaultValue=false&viewType=TABLE&areaType=BZN&atch=-
1016/j.apenergy.2016.12.065. false&dateTime.dateTime=01.01.2017+00:00%7CUTC%7CYEAR&dateTime.
[23] REFER; 2018. Available at: https://refer.upc.edu/ca (Accessed: 31 July 2018). endDateTime=01.01.2017+00:00%7CUTC%7CYEAR&are [accessed: 11 July
[24] Nordic TSOs. Unlocking flexibility Nordic TSO discussion paper on third-party ag- 2018].
gregators; 2017. doi: 10.1109/MPE.2016.2625218. [48] Koraki D, Strunz K. Wind and solar power integration in electricity markets and
[25] Ikäheimo J, Evens C, Kärkkäinen S. DER aggregator business: the Finnish case; distribution networks through service-centric virtual power plants. IEEE Trans
2010. p. 38. Power Syst 2017;33(1):1. https://doi.org/10.1109/TPWRS.2017.2710481.
[26] Paterakis NG, Erdinç O, Catalão JPS. An overview of demand response: Key-ele- [49] Convenant of mayors for climate & energy; 2018. Available at: https://www.eu-
ments and international experience. Renew Sustain Energy Rev 2017:871–91. mayors.eu/ [accessed: 6 August 2018].
https://doi.org/10.1016/j.rser.2016.11.167. [50] Barbero M. Montgat library consumption and temperatures. Mendeley Data 2020.
[27] Barcelona Energia; 2020. Available at: http://energia.barcelona/es/barcelona-en- https://doi.org/10.17632/ftwbtw9nmx.2.
ergia-la-electrica-publica-metropolitana (Accessed: 5 February 2020). [51] IEC. Electricity metering equipment (a.c.) – Particular requirements – Part 11:
[28] Good N, Ellis KA, Mancarella P. Review and classification of barriers and enablers of Electromechanical meters for active energy (classes 0, 5, 1 and 2); 2015.
demand response in the smart grid. Renew Sustain Energy Rev 2016. https://doi. [52] Klein SA, Duffie JA, Mitchell JC, Kummer JP, Thornton JW, Bradley DE, et al.
org/10.1016/j.rser.2017.01.043. Multizone building modelling with Type 56 and TRNBuild. Trnsys 16-a Transient
[29] Smart Energy Demand Coalition. Mapping demand response in Europe today. SEDC. System Simulation Program 2007;6:1–11.
Smart energy demand coalition; 2014, (April), p. 92. Available at: http://www. [53] Chang CY, Zhang W, Lian J, Kalsi K. Aggregated modeling and control of air con-
febeliec.be/web/infosession strategic demand reserve _ 16_5_2014/1011306087/ ditioning loads for demand response. IEEE Trans Power Syst 2013:1–10. https://
list1187970122/f1.html%5Cnhttp://sedc-coalition.eu/wp-content/uploads/2014/ doi.org/10.1109/ISGT.2013.6497895.
04/SEDC-Mapping_DR_In_Europe-2014-04111.pdf. [54] Canals Casals L, Barbero M, Corchero C. Reused second life batteries for aggregated
[30] Bondy DEM, Gehrke O, Thavlov A, Heussen K, Kosek AM, Bindner HW. Procedure demand response services. J Cleaner Prod 2019;212:99–108. https://doi.org/10.
for validation of aggregators providing demand response. 19th power systems 1016/j.jclepro.2018.12.005.
computation conference, PSCC 2016. 2016. doi: 10.1109/PSCC.2016.7540997. [55] Canals Casals Lluc, García Beatriz Amante, González Benítez MM. Aging model for
[31] Cappers P, Macdonald J, Goldman C. Market and policy barriers for demand re- re-used electric vehicle batteries in second life stationary applications. Project
sponse providing ancillary services in U.S. Markets, Lawrence Berkeley National management and engineering research. Lecture notes in management and industrial
Laboratory; 2013. doi: 10.1016/j.enpol.2013.08.003. engineering. Cham: Springer; 2017. p. 139–51. doi: https://doi-org.recursos.bi-
[32] Katz J. Linking meters and markets: Roles and incentives to support a flexible de- blioteca.upc.edu/10.1007/978-3-319-51859-6_10.
mand side. Utilities Policy 2014;31:74–84. https://doi.org/10.1016/j.jup.2014.08. [56] Restore; 2020. Available at: https://restore.energy/en/homepage [accessed: 5
003. February 2020].
[33] Rious Vincent, Perez Yannick, Roques Fabien. Which electricity market design to [57] Yuso; 2020. Available at: https://yuso.be/en/ [accessed: 5 February 2020].
encourage the development of demand response? Econ Anal Policy [58] SEAM group; 2020. Available at: http://www.seam-group.com/en/ [accessed: 5
2015;48:128–38. https://doi.org/10.1016/j.eap.2015.11.006. February 2020].
[34] Piette MA, Schetrit O, Kiliccote S, Cheung I, Li BZ. Costs to automate demand re- [59] Fortum; 2020. Available at: https://www.fortum.com/ [accessed: 5 February
sponse – taxonomy and results from field studies and programs; 2015. Available at: 2020].
https://gig.lbl.gov/sites/all/files/drrc_final_report_taxonomy.lbnl-1003924.pdf. [60] Smart Grid Energy; 2020. Available at: https://www.smartgridenergy.fr/ [accessed:
[35] ENTSO-E. ENTSOE’s transparency platform 2020; 2020. Available at: https:// 5 February 2020].
transparency.entsoe.eu/balancing/r2/balancingVolumesReservationPrice/show [61] Energy Pool; 2020. Available at: https://www.energy-pool.eu/en/technology/
(Accessed: 12 February 2020). [accessed: 5 February 2020].
[36] De Clercq B. Electricity security crisis in Belgium - Role of Elia system operator. [62] Voltalis; 2020. Available at: https://www.voltalis.com/business#header [accessed:
Elia, (January); 2015. 5 February 2020].
[37] Elia - Keeping the balance; 2019. [63] Open energy; 2020. Available at: http://www.openenergi.com/ [accessed: 5
[38] Energy Market Authority. National report 2013 to the agency for the cooperation of February 2020].
energy regulators and to the European commission. Finland; 2013. p. 1–68. [64] Kiwi power; 2020. Available at: https://www.kiwipowered.com/ [accessed: 5
Available at: http://www.ceer.eu/portal/page/portal/EER_HOME/EER_ February 2020].
PUBLICATIONS/NATIONAL_REPORTS/National Reporting 2013/NR_En/C13_NR_ [65] Flexitricity; 2020. Available at: https://www.flexitricity.com/en-gb/ [accessed: 5
Finland-EN.pdf. February 2020].

13