ABOUT SMARTEN ABOUT DNV

smartEn is the European business association DNV is an independent assurance and risk
for digital and decentralised energy solutions. management provider, operating in more than 100
Our members include innovators in services and countries. Through its broad experience and deep
technology for energy and data management, expertise DNV advances safety and sustainable
finance, and research. By taking an integrated performance, sets industry standards, and inspires
perspective on the interaction of demand and and invents solutions.
supply, our mission is to promote system efficiency,
encourage innovation and diversity, empower We provide assurance to the entire energy
energy consumers, and drive the decarbonisation value chain through our advisory, monitoring,
of the energy sector. verification, and certification services. As the
world’s leading resource of independent energy
For further information please visit: experts and technical advisors, we help industries
www.smarten.eu and governments to navigate the many complex,
interrelated transitions taking place globally and
regionally, in the energy industry. We are committed
to realizing the goals of the Paris Agreement, and
support our customers to transition faster to a
deeply decarbonized energy system.

For further information please visit:

www.dnv.com

Demand-side flexibility in the EU: Quantification of benefits in 2030, September 2022

AUTHORS:
Project Manager: Aurora Sáez Armenteros, Senior consultant (DNV)
Quality Assurance: Hans de Heer, Principal consultant (DNV)
Authoring team: Laura Fiorini, Consultant (DNV), María Miranda Castillo, Consultant (DNV),
Thijs Slot, Senior Consultant (DNV)

PROJECT MANAGEMENT AT SMARTEN:

Andrés Pinto-Bello Gómez, Head of Research and Projects (smartEn)

smartEn and DNV would like to thank smartEn members for their invaluable contributions and feedback
that made this report possible. In addition we thank Solarpower Europe for their cooperation facilitating
accurate datasets on the solar PV industry.

DESIGN:
Think Things Studio Barcelona
Executive Members

Regular Members

IKEA / INGKA

Associate Members
 Table of contents
EXECUTIVE SUMMARY __________________________________________ 6

1. INTRODUCTION _____________________________________________ 9
1.1 Context _______________________________________________________________________ 9
1.2 Purpose of this study _____________________________________________________________ 10
1.3 Approach ______________________________________________________________________ 10
1.4 Structure of this report ___________________________________________________________ 10

2. INPUT AND SET-UP OF THE MARKET MODEL ______________________ 11

2.1 CO2 emission target _____________________________________________________________ 12
2.2 Generation capacity mix __________________________________________________________ 12
2.3 Electricity demand _______________________________________________________________ 12
2.4 Reference scenario (no-DSF) _______________________________________________________ 15

3. DEMAND-SIDE FLEXIBILITY PROVIDED IN 2030 _____________________ 17

3.1 DSF available power and activated flexibility __________________________________________ 17
3.2 Industrial DSR __________________________________________________________________ 18
3.3 BESS – Behind the meter __________________________________________________________ 19
3.4 Smart charging _________________________________________________________________ 20
3.5 V2G ___________________________________________________________________________ 21
3.6 Residential electric heating ________________________________________________________ 22
3.7 Industrial electric heating _________________________________________________________ 23
3.8 Industrial Heating – CHP __________________________________________________________ 23
3.9 District heating – CHP _____________________________________________________________ 23

4. QUANTIFICATION OF DSF BENEFITS IN 2030 _______________________ 24

4.1 Socio-economic welfare and adequacy ______________________________________________ 24
4.2 System balancing ________________________________________________________________ 28
4.3 Grid infrastructure _______________________________________________________________ 29

5. QUANTIFICATION OF CONSUMER BENEFITS________________________ 30

5.1 Direct benefits __________________________________________________________________ 30
5.2 Indirect benefits ________________________________________________________________ 31
5.3 Stackability considerations ________________________________________________________ 32

6. KEY TAKE-AWAYS AND RECOMMENDATIONS _______________________ 33

APPENDIX A – DETAILED METHODOLOGY DESCRIPTION _____________ 35

1.1 Quantification of socio-economic welfare and adequacy benefits _________________________ 35
1.2 Quantification of system balancing benefits ___________________________________________ 44
1.3 Quantification of grid infrastructure benefits __________________________________________ 50
APPENDIX B – REFERENCES ____________________________________ 56
 EXECUTIVE SUMMARY
As the EU shifts away from the traditional energy
system and heads towards a decentralised, digitalised
Wholesale benefits
and decarbonised transition, new and smart solutions
In the year 2030 the model estimates a total of 164 GW
are required to manage the ever-increasing variable
upward flexible power1 and 130 GW of downward flexible
generation mix whilst maintaining affordability and security
power2. With an activation of 397 terawatt hours (TWh)
of supply. Demand-side flexibility (DSF), the ability of
of upward DSF and 340.5 TWh of downward DSF the
customers to change their consumption and generation
following results are found:
patterns based on external signals, is a crucial element
in achieving these goals.
€4.6 billion (5%) are saved due to lower costs to
generate electricity compared to a scenario without DSF.
With the current geopolitical events causing skyrocketing
energy prices and supply risk disruptions, the need to
The power system could serve all demand all year
empower end-users to play an active part in securing and
long, saving €9 billion on ‘lost load’ not served by the
decarbonising the EU energy system is gaining traction
available generation.
in EU strategies and policies. However, the activation
of consumers’ flexibility still faces regulatory barriers,
Renewable energy curtailment would be 15.5 TWh
notably due to the delayed implementation by Member
(61%) less, improving the economics of renewables and
States of the Electricity Market Design. Furthermore,
the availability of decarbonized electricity to consumers.
DSF still lacks visibility as a reliable, efficient and climate-
friendly solution because its potential has never been
37.5 million tonnes (Mt) would be saved in annual GHG
systematically quantified. This results in DSF being a
emission – i.e., 8%, nearly 84 kilos per capita, meaning
frequently overlooked solution in policy decisions,
that the power sector could exceed the ‘55% by 2030’
hindering its potential in accelerating the cost-efficient
target.
clean energy transition.

In order to fill this gap, smartEn – Smart Energy Europe,

Benefits for security of
commissioned an expert study from DNV to quantify supply
the potential benefits of a full deployment of DSF in the
EU by 2030. This DNV study intends to inform policy The modelling suggests that the energy system in
decisions on how to achieve a 55% GHG reduction by 2030 would lack at least 60 GW of generation capacity
2030 in a cost-efficient way for both the whole energy to ensure security of supply during the highest demand
system and consumers. peaks. Load shifting and load curtailment would allow
the system to maintain security of supply by fulfilling the
In a full-DSF activation scenario that unleashes the lack of generation capacity.
flexibility from buildings, electric vehicles and industry,
the following results are found: Enabling 60 GW of DSF would save €2.7 billion annually
compared to installing 60 GW of peak generation capacity.

Activating DSF technologies in European balancing

markets in 2030 could save a total of €262–690 million
across the EU27, a balancing energy cost saving of 43%
to 66%.

1. Upward DSF means increase of generation or decrease of demand.

6 2. Downward DSF means decrease of generation or increase of demand.
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Benefits for the distribution Direct benefits could lead to a potential cost reduction
for consumers of more than €71 billion (64%) per year
grid on electric consumption.

€11.1–29.1 billion would be saved in investment Over €300 billion in indirect annual benefits to people,
needs at EU 27 annually between 2023 and 2030. This communities, and businesses would result from reductions
represents between 27% to 80% of today’s forecasted in energy prices as a whole, generation capacity costs,
investment needs (between €253.1 billion and €282.5 investment needs for grid infrastructure, system balancing
billion between 2023 and 2030 in investments in low- costs, and carbon emissions.
and medium voltage distribution grids to integrate new
loads and RES capacity). This DNV study is a timely addition to the growing, but
still limited, corpus of detailed research into the potential
Benefits for consumers benefits of DSF to achieve the ultimate goals of providing
secure, accessible supplies of affordable clean energy to
The full deployment of DSF will translate into direct all consumers in the EU27. The findings serve as a clear
benefits for consumers with flexible assets, as well warning to not undervalue DSF given its huge potential
as indirect benefits to all customers through cheaper impact toward an efficient, clean electricity system.
electricity prices and lower grid costs:

What is Demand-side flexibility?

– the smartEn definition

“Demand-side flexibility” means the capability of any active customer to react to external signals and
adjust their energy generation and consumption in a dynamic time-dependent way, individually as well
as through aggregation.

Demand -side flexibility can be provided by smart decentralised energy resources, including demand
management, energy storage, and distributed renewable generation to support a more reliable,
sustainable and efficient energy system.

7
 Modelling and scenario-building

The input data and assumptions in the wholesale market simulation model are based largely on the Fit
for 55 objectives and REPowerEU. Drawing on these, DNV has defined inputs – divided into generation
mix, electricity demand, DSF technologies and CO2 emission target. DNV’s complex model assesses
monetary values of system-level savings and end-user benefits on the wholesale market from a full
activation of DSF.

The study explores two scenarios, ‘DSF’ and ‘no-DSF’, the latter providing a reference against which to
compare costs, benefits, emissions, and other outputs. Both have the same amount of flexible assets,
but in the DSF scenario these assets are fully price-elastic. In the no-DSF scenario, distributed flexible
assets are fully price-inelastic, though larger assets (electrolysers, front-of-meter batteries and central
generators) are fully price-elastic in both scenarios.

The estimated total benefit represents an order-of-magnitude value for the opportunity that could be
lost by failing to activate this level of flexibility.

DNV has performed calculations outside the model, using literature and simplified methodologies, to
estimate DSF benefits for adequacy, balancing and grid infrastructure costs.

CONSIDERATIONS:

The model developed and applied for this study, Although DSF investments are substantially lower
and its results, are constrained by limitations of data than the other technologies mentioned, it is unclear
quality, comparability, and availability across Member how much the full potential will develop on its own
States in the EU 27. – assuming all regulatory barriers are removed – or
whether additional incentives are needed, as we have
DNV also points to a lack of studies on the seen and continue to see for batteries, electrolysers,
quantification of infrastructural benefits of DSF, and renewables.
suggesting that this is a signal for relevant stakeholders
to further investigate the topic. The model did not take into account the positive
energy efficiency impact of DSF activations, nor
The no-DSF scenario is an unrealistic one given that potential savings in TSO redispatch costs and TSO grid
flexibility is already activated to varying degrees in reinforcement costs.
several EU member states. The DNV study explains in
detail how this scenario has been carefully constructed Benefits within the four segments (wholesale,
to allow a meaningful quantification of the potential adequacy, balancing and distribution grids) have been
value of DSF to the energy system. calculated separately. The total DSF benefits are lower
than the sum of the benefits per segment, due to the
The model is an energy-only one, based on marginal close interaction of these four segments.
costs, in line with the current market functioning. The
model does not consider capital expenditure costs for Gas prices considered in the model are moderate
generation assets, batteries, electrolysers or DSF, except compared to the exceptional 2022 levels.
for the quantification of the security-of-supply benefits.

8
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

1. INTRODUCTION
The use of Demand-Side Flexibility (DSF) in our European power system has been advocated for many years as a
crucial tool that empowers consumers to help integrate renewable energy sources in the energy system, and to
increase the overall efficiency of system development and day-to-day operations. More DSF deployment should
result in having more flexibility in the system at times of peak electricity demand, thereby reducing the need for
more expensive sources of power generation to remain available and be dispatched and thus reducing the overall
prices to consumers. Additionally, DSF could help to reduce the need for power grid investments, since the availability
and deployment of more flexibility on the demand side can help to reduce strain on critical components and/or
connections in the grid. This is especially so in light of the fundamental system change in which fully controllable,
fossil-based power plants are rapidly being replaced by renewable energy sources, whose production is largely
determined by weather conditions.

Consumers, however, largely consider electricity to be a commodity that they consume whenever they need it. Being
more flexible in the way they consume their power is typically not something they are concerned about, largely
because most consumers pay a fixed price per kilowatt-hour (kWh) to their retail supplier.

To advance insights on what DSF can mean for the (transition of the) European power system, smartEn commissioned
a study from DNV, aimed at quantitatively assessing the potential benefits of a full deployment of DSF in the EU 27
Member States by the year 2030.

1.1. Context
Although DSF has been earmarked as an important In the context of the Fit for 55 package and the associated
source of power system flexibility, its actual contribution efforts that Member States are undertaking to achieve
to current system operations and the power market 55% greenhouse gas (GHG) emission reductions by
remains largely unknown. Even though EU directives and 2030, DSF may have a key role in providing the flexibility
regulations aim to stimulate a more widespread adoption needed to allow for this objective to become a reality.
and market participation, overall contributions (responding In addition, recent events such as the war in Ukraine,
to system needs that are reflected through market prices) and aggravated climate concerns, are driving a (much)
appear to remain limited. Ahead of this study into EU- quicker phase-out of the use of (Russian) gas and a
wide benefits, there are no (public) sources available ramp-up of emission reduction targets. These updated
that have investigated overall availability, deployment goals are incorporated into the European Commission’s
and/or pricing of DSF across EU 27 nations. Some of the REPowerEU Communication and the emergency electricity
available publications provide more detailed insight into market design interventions.3 These updated strategies
the possible contributions of specific forms of demand also include a vast expansion of distributed flexible assets
response, benefits for specific countries, or contributions and requirements for their activation, intended to be
to cost savings in specific areas (e.g. grids in a specific able to serve even more of Europe’s energy needs with
country). renewably generated power.

3. See: REPowerEU: affordable, secure and sustainable energy for Europe | European Commission (europa.eu)
9
 1.2. Purpose of this study
The purpose of this study is to assess the potential benefits market simulations in a ‘DSF scenario’ for 2030 in DNVs
of DSF to the European power system in the year 2030, European Market Model, an economic dispatch simulation
where all renewable energy sources (RES) targets and model to simulate power markets. Results are compared
55% GHG reduction are achieved, assuming that DSF against a ‘no-DSF scenario’ that assumes no flexibility to
can access all markets throughout the EU 27. The project be available from the DSF-technologies whose impacts
team has developed and executed an approach to be are assessed in this report (see 2.3.1 for a list of the
able to identify: selected DSF technologies). This comparison shows
differences between the two scenarios in terms of reaching
emission reduction targets and the costs for generation
The DSF capacities that can be available in 2030
and consumption. It should be noted that the no-DSF
scenario is not realistic and is not a counterfactual, but
rather a reference to calculate the maximum achievable
The DSF volumes that would be utilised to optimise potential.
the wholesale market in 2030

Benefits for the balancing market and grid development

are approximated based upon deductions from various
The potential impact that DSF activations have on sources, e.g. departing from specific grid savings in a
wholesale, adequacy, system balancing and grid
infrastructure single Member State. Although this may be assessed in
much more detail later (e.g. by assessing the specific grid
situation and challenges to overcome in each Member
State and at every grid level), the project team expected
How system-wide DSF benefits translate into benefits in wholesale trading and adequacy to be much
benefits for consumers
larger due to sheer trading volumes across all EU 27
Member States. Therefore, a thorough analysis has first
Although the study addresses all areas – wholesale been performed on these two aspects.
market, system balancing, and infrastructure – the focus
of the approach and modelling is on wholesale because Further details about our approach and its limitations
the potential impact was foreseen to be much larger. are provided in chapter 2 and Appendix A.
Throughout the study, ‘wholesale’ is defined as forward,
day-ahead and intraday electricity markets (explicitly 1.4. Structure of this report
excluding balancing markets / ancillary services).
Chapter 2 provides further details of the approach
1.3. Approach regarding wholesale market simulations in two different
scenarios (DSF and no-DSF). Chapter 3 discusses the
Specific industry insights about technology developments results of the market simulations regarding the different
and policy targets are combined with those from available DSF-technologies. Overall benefits in terms of wholesale
public sources on the availability and impacts of DSF- benefits, emission reductions, generation adequacy,
technologies, such as reports by Eurelectric, ENTSO-E balancing and grid development across the EU 27 are
and individual TSOs such as RTE (France), Elia (Belgium), the focus of chapter 4. Chapter 5 assesses the resulting
E-Bridge and DNV. These inputs and datasets, and consumer benefits. Finally, chapter 6 provides an overview
contributions by smartEn members, made up the inputs of DNVs main conclusions and recommendations.
for this study.
Appendix A provides further detail about the approaches
Electricity wholesale market benefits are assessed by to quantify the impacts of DSF in the various areas
using these (additional) DSF options as input to wholesale (wholesale, balancing and grid development).

10
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

INPUT AND SET-UP OF

2. THE MARKET MODEL

The main focus of this study is the analysis of the potential wholesale benefits of DSF use in the European power
system in 2030. This focus was based on the assumption that, out of the four main value drivers (wholesale, adequacy,
balancing and grid infrastructure), the first inhibits the highest potential for DSF to valorise – an assumption that is
confirmed by the outcome of this study. To quantify these potential wholesale benefits of DSF use, DNV modelled
the European power market considering the different DSF options available in 2030.

DNV has used its European Market Model, a fundamental market model that simulates the day-ahead spot price by
optimising the unit commitment and economic dispatch of electricity generation. The simulations are performed
on an hourly time-resolution containing a detailed representation of generation, commodity prices and demand
for all bidding zones in EU 27 Member States, based on the following modelling assumptions:

Generation capacities are modelled on an individual against certain constraints within the model (see Appendix
basis with detailed techno-economic characteristics A – section 1.1.2) – e.g. electric vehicles (EVs) need to
such as, but not limited to, heat rates, ramping ability, be charged by a certain volume within a specified period
minimum stable level, fuel cost, other variable operating (e.g. during the night or within one week).
costs, maintenance and forced outage rates, etc.
The model set-up assumes that all flexible demand and
Renewable generation takes volatility into account generation, both front and behind-the-meter, is exposed
through the use of historical or re-analysed time series to the market.
of, for example, data on wind speed and solar irradiation
for different locations. These profiles take geographical The commodity prices are set at a “normal’ level
correlation into account. excluding exceptional situations such the Ukrainian war as
well as the low availability of the French nuclear portfolio.
Market exchanges between countries (i.e. bidding
zones) are defined based on Net Transfer Capacities. The Network tariffs and taxation are not included in the
increase in available transmission capacity is based on model. As a consequence, there is no explicit optimisation
available projections announced by individual TSOs and/ of (collective ) self-consumption. In practice, this effect
or ENTSO-E. Transmission and distribution constraints is implicit to the system behaviour. For example, if a
within bidding zones are not modelled. residential customer has both rooftop PV and a battery,
then the battery (being exposed to market prices) will
The demand consists of an hourly fixed demand profile, typically charge when high amounts of PV energy are
flexible demand-side management components and other produced – yielding a similar result as self-consumption
flexible load originated by front-of-the-meter applications optimisation.
such as utility-scale batteries. Flexible demand is optimised

Finally, the input data and assumptions included in the model are focused on the Fit for 55 objectives (European
Commission , 2021) and REPowerEU Communication (European Commission, 2022). Considering these guidelines,
DNV has defined the inputs to the European Market Model, which can be divided into generation mix, electricity
demand, DSF technologies and carbon dioxide (CO2) emission target.

11
 2.1. CO2 emission target
Fit for 55 establishes the target of reducing net GHG emissions for all Member States by at least 55% by 2030,
compared with 1990 levels. DNV translated this target into the equivalent of 410 million tonnes (Mt) of CO2 emissions
in the EU power sector by 2030. This CO2 emission target is based on the result of DNV’s Energy Transition Outlook
model (DNV, 2021). This target was used as a benchmark to evaluate the ability of the 2030 scenario, and DSF in
particular, to help reach the Fit for 55 target.

2.2. Generation capacity mix

Considering the EU 27 Member States, the generation capacity mix in 2030 is characterised mainly by a high share
of renewables, reaching 75% renewable installed capacity. Thermal installed capacity is reduced to less than 25% of
the generation portfolio, as seen in Figure 1. Gas-fired generation is the main thermal source4, while coal and lignite
generation are significantly reduced. The development of generation installed capacities are in line with the targets
presented in REPowerEU and Fit for 55 for an increase in renewable energy production, an increment of renewable
hydrogen production, and a consequent decrease in GHG emissions. In particular, installed solar PV capacity (front
and behind the meter) in EU 27 reaches 600 GW by 2030 as defined in the REPowerEU Plan. Moreover, offshore
wind capacities in the North-Sea are increased based on the latest targets agreed by Belgium, Denmark, Germany
and the Netherlands, reaching 65 GW by 2030 (BMWK, 2022).
Figure 1 – Installed generation capacity EU 27 in 2030 (GW)

2030

0 200 400 600 800 1000 1200 1400 1600

Nuclear Gas Coal Lignite Oil Biomass Waste

Hydro Solar Wind offshore Wind onshore

Figure 1 – Installed generation capacity EU 27 in 2030 (GW)

2.3. Electricity demand

The electricity demand is considered in the analysis by differentiating the traditional demand for electricity from
the additional demand due to the electrification of passenger transport and heating, and from electrolysers for
power-to-hydrogen conversion.

Traditional demand encompasses, for example, household, commercial and industrial power demand, categories
already considered nowadays. This segment of demand reaches 2858 TWh across EU 27 Member States in 2030.

Electrification of passenger transport5 is driven by support schemes and by technological and infrastructure
developments and expected cost degression. This creates demand for 151 TWh of electricity in 2030.

4. Gas prices are assumed to decrease by 2030. The model considers a gas price of 25.3 €/MWhth in 2030.
12 5. In our model this is limited to passenger EVs.
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Electrification of heating consists of both space heating and industrial heating, accounting for 510 TWh by 2030.6

Power-to-hydrogen entails the electricity demand required by electrolysers. The electrolysers’ demand increases
significantly in 2030 to reach the targeted 10 Mt of renewable hydrogen production in Europe, based on the REPowerEU
Communication. Therefore, using European Commission assumptions and according to DNV calculations, 562 TWh
of electricity consumption for hydrogen production is expected in 2030.7
Figure 2 – Electricity demand EU 27 in 2030 (TWh)

2030

0 200 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500

Traditional demand Electric vehicles Electric heating Electrolyser power consumption

Figure 2 – Electricity demand EU 27 in 2030 (TWh)

2.3.1. DSF TECHNOLOGIES

To quantify the potential benefits of demand-side Other DSF technologies with relevant potential were not
flexibility, DNV has modelled DSF technologies as part included in the list for this study, due to unavailability
of the European power system used in the analysis. of sufficient data to assess their availability throughout
Based on the technology characteristics and high EU 27 by 2030. These include, for example, district
flexible power potential, DNV included the following cooling, residential cooling, Joule effect electric heating,
DSF technologies: and residential electric boilers. To mitigate the risk of
overestimating total DSF capacities that are available
Smart charging to the power system in 2030, DNV has decided to not
incorporate these technologies into the study.
Vehicle-to-grid
Finally, behind-the-meter solar PV is also considered in
Behind-the-meter (BTM) batteries the model. However, it is not modelled as a controllable
asset but as non-curtailable PV generation.
Industrial demand-side response (DSR)

Residential space electric heating

Industrial electric heating

District heating – combined heat and power (CHP)

Industrial heating – CHP

7. European Commission assumptions are a utilisation factor of 43% and electrolyser efficiency of 70%.
6. Water heating is not included. 13
 Smart charging and vehicle-to-grid Residential electric heating

Electric vehicles are included in the analysis and considered In the model, all flexible residential electric heating is
as DSF technologies that can provide flexibility by shifting assumed to be provided by heat pumps and is considered
their load. Smart charging is modelled by optimising the a shiftable load within 12-hour periods. Therefore, the
total daily EV demand when EVs are connected to the grid. half-daily load required for residential electric heating
For a detailed description of smart-charging modelling is met, but the hours when the consumption take place
and references refer to Appendix A – section 1.1.2.3. can shift overtime. For detailed description of residential
electric heating modelling and references refer to Appendix
A total of around 60 million EVs by 2030 are included for A – section 1.1.2.5. In order to represent this load, DNV
the 27 Member States. has considered a total space heating electricity demand
of 449 TWh by 2030.
Additionally, DNV assumed that 30% of the EV chargers are
enabled for bidirectional charging (European Commission, Industrial electric heating8
Directorate-General for Energy, 2022). Bidirectional ch
arging or vehicle-to-grid (V2G) enables the EV battery Industrial electric heating (e-boilers) is considered a
to feed in to the grid as well as charging. As such, V2G is curtailable load. The load would curtail above a given
modelled as behind-the-meter battery whose charging electricity price based on the cost that the industrial plant
and discharging is limited by the EVs that are connected would incur when switching off the e-boiler. For detailed
to the grid. For detailed description of V2G modelling and description of industrial electric heating modelling and
references refer to Appendix A – section 1.1.2.4. references refer to Appendix A – section 1.1.2.6. DNV
has considered 7 GW of industrial heating load across
Behind-the-meter batteries all Member States.

BTM batteries provide flexibility through charging and District heating – CHP
discharging daily when prices show a sufficient spread to
cover their efficiency and operational costs. For detailed District heating – CHP is considered as aggregated CHP
description of Battery Energy Storage System (BESS) plants with a daily generation requirement. This DSF
modelling and references refer to Appendix A - section technology behaves as a generator that can always deviate
1.1.2.2. In this study, DNV has considered a total of 10.9 upwards from their daily generation requirement, but can
GW of BTM batteries in the EU 27. only deviate downwards when it is more optimal to pay
a penalty than to generate (the penalty price is based on
Industrial demand-side response the alternative cost for heating). For detailed description
of CHP modelling and references refer to Appendix A –
Industrial DSR is considered a curtailable load. The load section 1.1.2.8. DNV has included a total of 56 GW of CHP
would curtail above a given electricity price that varies district heating capacity in EU 27 by 2030.
per category and is based on the cost that the industrial
plant would incur when interrupting its operation. For Industrial heating – CHP
detailed description of industrial DSR modelling and
references refer to Appendix A – section 1.1.2.1. A total Industrial heating – CHP is represented as CHP plants with
of 21.7 GW of industrial DSR capacity is included across specific daily generation requirements. This technology
all Member States. is modelled following the same logic as district heating –
CHP, with different generation requirements and penalties.
For detailed description of CHP modelling and references
refer to Appendix A – section 1.1.2.7. DNV has included a
total of 19 GW of industrial CHP capacity by 2030.

8. This category is limited to heating by e-boilers due to data unavailability at European level on other technologies such as heat pumps.
14
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

2.3.2. OTHER FLEXIBLE TECHNOLOGIES

The following technologies are included in the model as flexible technologies but are not considered as demand-side
flexibility because they are not placed behind the meter. The flexibility that these technologies provide is therefore not
included in the DSF capacities and their possible benefits that this study sets out to quantify. At the same time, since
these technologies are actively participating in the electricity market, they do influence the outcome of this study.

Grid-connected storage Electrolysers

Grid-connected batteries provide flexibility through Electrolyser consumption is considered as a flexible load.
charging and discharging daily when prices show a The load would curtail above a given electricity price
sufficient spread to cover their efficiency and operational based on the cost of hydrogen, this is estimated at 86.2
costs. For detailed description of BESS modelling and €/MWh. For detailed description of electrolyser modelling
references refer to Appendix A – section 1.1.2.2. In this and references refer to Appendix A – section 1.1.2.9.
study, DNV has considered a total of 15.5 GW of front- Installed capacities and annual consumptions are based
of-the meter batteries in EU 27. on the REPowerEU Communication’s indicative targets
of 10 mt renewable hydrogen annual production, which
may translate to up to 149 GW in total for all Member
States by 2030.9

2.4. Reference scenario (no-DSF)

To evaluate the outcomes of the model and derive the potential DSF contribution, it is necessary to set a reference
against to which to compare costs, emissions, and other parameters.

Lack of data on the DSF utilised today, and on DSF utilisation prospects towards 2030, makes it impossible to quantify
the counterfactual system costs. Therefore, DNV modelled a 2030 scenario in which the demand side is not flexible.
Realising that demand-side technologies are to some extent – and in some Members States more than others –
already actively participating in the electricity market, DNV acknowledges that the reference scenario is not a realistic
one. The main purpose is to quantify the total potential benefit of DSF for the considered technologies. If there are
still certain barriers in 2030 to DSF responding to external price signals, a certain share of this potential will not be
achieved. The total benefit provides an order-of-magnitude value of this missed opportunity (albeit an upper bound).

The no-DSF scenario considers the same amount of DSF technologies (e.g. the same amount of EVs), yet all DSF
technologies have a fixed demand/generation profile and therefore show no price responsiveness – i.e. they provide
no flexibility to the system.

The no-DSF scenario does include the same amount of other flexibility sources that are not considered to be DSF,
mainly electrolysers and front-of-the-meter storage. These flexible resources are price responsive and are treated
as such in the no-DSF scenario.

Throughout the report the reference scenario in which DSF technologies are not flexible is referred to as the ‘no-
DSF’ scenario; and the scenario in which DSF technologies are flexible is referred as the ‘DSF scenario’.
The table 2.1 provides a summary of the implementation of the DSF and no-DSF scenarios. The detailed description
of the modelling of both scenarios is included in Appendix A – section 1.1.2.

9. This capacity value was calculated based on the targeted renewable hydrogen production in REPowerEU and its reported assumptions.
15
 Table 2.1 – DSF and no-DSF scenario implementation

DSF scenario No-DSF scenario

Industrial DSR capacity is There is no industrial DSR capacity, all

Industrial DSR
price responsive is modelled as fixed traditional load

BESS capacity provides flexibility BESS systems behind the meter do not
BESS – behind the meter
feed-in or off-take electricity.

EV charging is optimised against EV charges following a fixed hourly

Smart charging prices profile

V2G V2G capacity provides flexibility There is no V2G

Residential electric heating is Electric heating demand follows a

Residential electric heating
optimised against prices fixed hourly profile

CHPs can increase/decrease

Industrial electric heating – CHP generation to provide flexibility
CHPs follow a fixed generation profile

CHPs can increase/decrease

District heating – CHP generation to provide flexibility
CHPs follow a fixed generation profile

BESS – front of the meter BESS capacity provides flexibility BESS capacity provides flexibility

Electrolysers are price responsive, Electrolysers are price responsive,

Electrolysers i.e. flexible i.e. flexible

16
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

DEMAND-SIDE FLEXIBILITY
3. PROVIDED IN 2030
This section presents the results on quantification of DSF in 2030, based on wholesale optimisation. First, the total
flexible power and activated flexibility is reported at EU 27 level. Then, the modelling of flexibility and its quantification
approach is detailed per technology.

3.1. DSF available power and activated flexibility

DSF available power

Based on the inputs and outputs of the market model, there is a total of 164 GW upward flexible power and 130
GW of downward flexible power in 2030. Considering 752 GW of peak demand of the EU 27 system in 2030, the
upward and downward flexible power represents about 22% and 17% of the peak demand, respectively.

Table 3.1 summarises the available flexible power per technology, on average in 2030. It shows both upward flexibility
(increasing generation or reducing demand) and downward flexibility (decreasing generation or increasing demand).

It should be noted that the way the flexible power is calculated varies per technology. This is because the technologies
are flexible in different ways, e.g. smart charging available power is not constant but depends on the number of
cars connected to chargers that have available battery capacity.

Table 3.1 – Available flexible power

Upward flexible Downward flexible

Technology power [MW] power [MW]

Industrial DSR 21,731 0

BESS Behind the meter 10,850 10,850

Smart charging 48,704 16,295

V2G 25,594 25,594

Residential electric heating 32,841 73,385

Industrial electric heating 7,082 0

Industrial heating – CHP 6,355 482

District heating – CHP 10,581 3,500

Total 163,738 130,106

17
 Activated flexibility

DNV calculated the amount of flexibility that was activated as a result of the 2030 electricity (wholesale) market
simulation that optimises the system behaviour. The results indicate that a total of 397 TWh and 340 TWh of upward
flexibility and downward flexibility, respectively, are activated in 2030 within EU 27. Considering 4,081 TWh of total
demand in 2030, the upward and downward activated flexibility corresponds to about 10% and 8% of the total
demand, respectively.

Table 3.2 summarises the total activated flexibility broken down per DSF technology. The results show that the
largest part of activated flexibility in both directions is supplied by residential electric heating, followed by EVs,
CHPs supplying district heating, and V2G.

Table 3.2 – Total activated flexibility

Upward flexibility Downward flexibility

Technology [GWh] [GWh]

Industrial DSR 1,071 -

BESS Behind the meter 637 871

Smart charging 106,286 106,266

V2G 21,009 23,764

Residential electric heating 195,532 195,532

Industrial electric heating 141 -

Industrial heating – CHP 12,697 12

District heating – CHP 59,601 14,032

Total 396,974 340,477

The next subsections include a more detailed description of the model results and the calculation of flexible power
and activated flexibility per technology.

3.2. Industrial DSR To illustrate the realised capacity provided by industrial

DSR, the load duration curve of industrial DSR activation
All energy generated by industrial DSR assets is considered is included below (Figure 3). The load duration curve
upward flexibility as it corresponds to a reduction in shows the duration (% of the hours in a year; x-axis) of
demand. When the entire cheaper generation capacity deployment of a certain load level (MW; y-axis) throughout
is already fully committed, industrial DSR acts as a last all the 8,760 hours of 2030. It can be observed that
resort to prevent unserved load. The results of the model the cheapest industrial DSR option (i.e. 207 €/MWh) is
show that the average price of activated industrial DSR the most frequently activated, about 500 hours in one
is below 550 €/MWh in all Member States except in two year for the whole EU 27, whereas the most expensive
Member States, where more expensive categories are category (i.e. 1,097 €/MWh) is activated for less than 1%
also activated sometimes. of the time. Overall, the total activated upward flexibility
is 1,071 GWh.

18
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Figure 3 – Industrial DSR power duration curve

3.3. BESS – Behind the meter

All energy either supplied or consumed by behind-the- One can see in Figure 4 an illustrative example of the
meter batteries is considered upward or downward operation of behind-the-meter BESS in one week of
flexibility, respectively. The total activated upward flexibility 2030 in Germany. The batteries follow the market price:
is 637 GWh, whereas the total activated downward when the price variations are large enough to overcome
flexibility is 871 GWh. The difference between upward and operating inefficiencies and costs, the batteries charge
downward flexibility is due to charging and discharging if the price has dropped, or discharge if the price has
inefficiencies. increased. Note that this optimisation is done for one
day look-ahead, therefore sufficient price spread need
For this technology, the flexible power is calculated to be fulfilled during the same day.
the installed capacity. This capacity is available in both
directions, charge and discharge.

Figure 4 – Example of behind-the-meter BESS (dis)charging during week 29 in Germany in 2030

19
 3.4. Smart charging
The flexibility provided by EVs is calculated as the difference is calculated as the average deviation from the reference
between the optimised EV load given by the simulation charging profile respectively.
results and a fixed EV charging profile used as reference
in the no-DSF scenario, while respecting the total daily To show the difference between the optimised smart
load requirement. Upward flexibility is provided when charging and a non-flexible EV charging behaviour, Figure
the optimised EV load is below the reference charging 5 includes the resulting profile from the market model
profile, whereas downward flexibility is provided when against the non-flexible EV profile. The results correspond
the optimised EV load is above the reference charging to EVs during week 4 of 2030 in Spain.
profile. As such, the upward and downward flexible power

Figure 5 – Comparison between optimised EV charging and reference

EV charging profile of Spanish EVs during week 4 of 2030

Downward activated flexibility absolute values are the same as the EV daily consumption per Member State is a
given input to the optimisation challenge. Overall, the total activated flexibility is 106.3 TWh in both directions.

Figure 6 illustrates an example of the behaviour of EVs in week 30 in Spain resulting from the market model. The
figure shows how the daily EV load is optimised against market prices while respecting the EV capacity available
each hour. As the daily EV fleet consumption is a required input of the model, the EVs charge every day, preferably
during hours of low prices. Additionally, the hourly maximum charging power is capped by the availability profile,
which is defined as the hourly charging capacity connected at private and office charging points.

20
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Figure 6 – Example of EVs charging in Spain during week 30 of 2030

3.5. V2G
All energy either supplied or consumed by the EVs via V2G is considered upward or downward flexibility, respectively.
For this technology, the flexible power corresponds to the total charging point capacity with V2G capability (i.e.
30%) multiplied by the average utilisation factor (26.1%). Overall, the total upward flexibility is 21 TWh and the total
downward flexibility is 23.7 TWh. Load values are 12% higher due to charging/discharging inefficiencies.

One can see an illustrative example of the behaviour of V2G in a week in the Netherlands in 2030 in Figure 7. V2G
follows the market prices: when the price variations are large enough to overcome operating inefficiencies and
costs, the batteries charge if the price has dropped, or discharge if the price has increased, keeping the number of
daily cycles below or equal to 2. Additionally, the max charge or discharge power is constrained by the availability
profile (both consumption and feed-in) of charging points, as shown in Figure 8.

Figure 7 – Example of V2G behaviour in the Netherlands during a week of 2030

21
 Figure 8 – Example of V2G behaviour in the Netherlands in the 2030 DSF scenario
within feed-in and load availability profiles

3.6. Residential electric heating

The flexibility provided by residential electric heating is defined as the difference between the optimised load
given by the simulation results and the hourly profile used as reference in the no-DSF scenario. Upward flexibility
is provided when the optimised load is below the reference profile, whereas downward flexibility is provided when
the optimised load is above the reference profile. As such, the upward and downward flexible power is determined
by the average respective deviation from the reference hourly profile.

An Illustrative comparison between the ‘optimised’ load and the reference profile of German residential electric
heating during week 11 of 2030 in the DSF scenario is shown in Figure 9. One can see that the load is preferably
supplied in hours with low power prices, while fulfilling the 12-hour consumption requirement.

For all EU 27, the total activated flexibility is 195.5 TWh in both directions.

Figure 9 – Example of residential electric heating in Germany during week 11 of 2030

22
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Several studies show that demand response, when applied to (residential) electric heating, also leads to energy
efficiency; the total energy used will be lower than the counterfactual (no DSF). This is an important upside, as it
leads to direct customer savings, lower demand (thus lower market prices), and carbon savings. The main reason
for not including this in our market modelling and quantification is the lack of empirical data that is both relevant
for the technology considered (heat pump) and representative for EU 27.

3.7. Industrial electric heating

The flexibility provided by industrial electric heating is An illustrative comparison between the optimised load
determined by the difference between the non-flexible and the reference profile of Finnish industrial electric
industrial electric heating demand (fixed) profile and heating during week 1 of 2030 is shown in Figure 10.
the optimised demand profile given by the simulation One can see that the load is curtailed when the power
results. Upward flexibility is therefore provided when price rises above 500 €/MWh.
the optimised demand is below the constant non-flexible
profile. As such, the upward flexible power is calculated Overall, the total activated upward flexibility for EU 27
as the total installed capacity of flexible industrial load. in 2030 is 140.7 GWh.

Figure 10 – Example of industrial electric heating behaviour in Finland in week 1 of 2030

3.8. Industrial heating – CHP 3.9. District heating – CHP

Industrial heating – CHP is modelled as a CHP plant with District heating – CHP is modelled as aggregated CHP
a power generation profile set at 100% of max capacity plants with a daily generation requirement which is linked
between 05:00 and 20:00 hours, and at 65% otherwise. to the daily heat demand. CHP units are defined as part
Upward flexibility is, therefore, provided when the of a heating area for which they must fulfil certain level
generation is above the hourly target production, whereas of generation per day, which represents the required heat
downward flexibility is provided when the generation is consumption. Therefore, flexible power and flexibility
below the hourly target production. As such, the upward activation are calculated following the same method as
flexible power is calculated as the average deviation above for industrial heating, namely CHPs.
the hourly target production, whereas the downward
flexible power is defined as the average deviation below The total activated upward and downward flexibility are
the hourly target production. 59.6 TWh and 14.0 TWh, respectively.

Overall, the total activated upward and downward

flexibility are 12.7 TWh and 12 GWh, respectively.

23
 QUANTIFICATION OF
4. DSF BENEFITS IN 2030
4.1. Wholesale markets and adequacy
The underlying assumption of this study is that wholesale market benefits outweigh the potential DSF benefits such
as balancing or infrastructure savings, hence DNV’s main focus was on the modelling and methodology to derive
these market benefits. This section and the following ones prove that this assumption is correct and wholesale
market benefits are significantly higher than the rest.

To quantify the effect of the optimal DSF deployment on wholesale markets in EU 27, DNV has modelled and simulated
the European Power Market for the year 2030 (for the DSF and the no-DSF scenario), as described in section 2.10 The
benefits were subsequently quantified by different metrics as the difference between the results for both scenarios:

Cost to generate:

These are the costs that generators/storage incur to cover the system demand. These include fuel costs, variable
operation and maintenance (VOM) costs, start and shutdown costs, emissions costs and penalties.11

Loss of load:

This is the total amount of load that has not been served by the available generation. The cost used to monetise
the loss of load is set to 3,500 €/MWh. The actual valuation will strongly depend on the context, in particular the
degree of acceptability of load shedding and load curtailment as a mean of system adequacy.

Renewable energy curtailment:

This includes the curtailment of renewable energy generators for economic reasons and interconnector congestion
considerations.

Greenhouse gas emissions:

These are the total emissions by generators to cover the power demand.

Cost to serve load:

This is the total price that load needs to pay for their electricity intake. The price for the load is considered as the
day-ahead hourly spot price.

10. More details on the methodology and limitations of his methodology can be found in Appendix A – section 1.1
24 11. Penalties apply when a generator or a demand unit violates a constraint, e.g. a district heating CHP would pay a penalty when it generates less than the required production.
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

The results (Table 4.1) indicate that, in 2030, with the Cost to serve load benefits are significantly higher than the
activation of 397 TWh upward DSF and 340.5 TWh of rest of the parameters. This highlights the considerable
downward DSF: impact that load curtailment and load shifting have on
the generation mix for each moment. DSF avoids the
The cost to serve load, i.e. the cost that consumers creation of high peaks where very expensive (and price
pay for their electricity consumption, is around €301.5 setting) generators are needed, and absorbs the energy in
billion (48%) less than in the no-DSF scenario; the case of a generation surplus and relatively low prices.
Therefore, it can be observed that even if the generator
The system meets and even exceeds the necessary costs are only 5% less, the lower utilisation of expensive
emission reduction in the power sector to fulfil the 55% generators makes a tremendous impact on the final cost
GHG reduction, whereas the no-DSF scenario does not to load (nearly 50%).12
achieve the target;

The emissions are 37.5 million tonnes (8%) less than

in the no-DSF system;

Costs to generate are €4.6 billion (5%) less than in a

no-DSF system;

The system can serve all demand throughout the year,

whereas the no-DSF system cannot serve all the 2,054
GWh of load in 2030. Therefore, the DSF system saves
€9 billion on value of lost load; and

Renewable energy curtailment is 15.5 TWh (61%) less

than in the no-DSF scenario.

Table 4.1 – Wholesale benefits of DSF

Benefits (DSF scenario vs. Relative to no-DSF

Year 2030 – EU 27 no-DSF scenario) scenario

Cost to serve load - €301.5 bn - 48%

Emissions - 37.5 Mt - 8%

Cost to generate - €4.6 bn - 5%

Value of loss of load - €8.97 bn - 100%

Renewable energy curtailment - 15.5 TWh - 61%

12. It should be noted that if the security of supply was met in the no-DSF scenario, there would have been a (lower) different outcome on benefits.
25
 The strong impact of DSF on the (residual) demand curve leads to a shift from generator’s surplus to consumer’s
surplus. When comparing the DSF to the no-DSF scenario, the larger part of the savings on the cost to serve can
therefore be attributed to the reduction of generator’s margins (which are, at least partly, required to cover the
initial investments). This impact is further specified in Table 4.2, which shows the total generation margins per type
of generation technology, for both scenarios. Since the no-DSF scenario is not realistic, these figures do not fully
describe the impact of full DSF deployment on the investment climate for different generation technologies. It does
indicate, however, which technologies are affected most when deploying DSF to its full potential:

Profitability of fossil fuel plants is very strongly impacted as both their running hours and market price volatility
are strongly reduced;

Profitability of nuclear plants is strongly impacted mainly due to reduced market price volatility, however nuclear
can benefit from the floor price already noticeable in the no-DSF scenario, see also Figure 11;

Profitability of renewables is somewhat impacted. However, the large amount of electrolysers and front-of-meter
storage already creates a floor price in the no-DSF scenario, and therefore the impact for renewables seems limited;

Profitability of biomass is strongly impacted, although less than fossil-fuel generators.

Table 4.2 – Impact of DSF on generator types

Generation
technology type
Fossil fuel Nuclear Renewable Biomass

Margin
Billion € Revenues Costs Margin Revenues Costs Margin Revenues Costs Margin Revenues Costs

No-DSF scenario 175,698 83,733 91,965 101,570 4,527 97,043 218,151 4,924 213,227 29,425 3,802 25,624

DSF scenario 86,359 77,369 8,990 48,648 4,594 44,054 158,553 4,440 154,113 11,577 4,201 7,376

Relative change -90% -55% -28% -71%

The impact of a full deployment of DSF on market price volatility is further demonstrated by the price duration
curves below.

Figure 10 – Price duration curves for both scenarios, full (left) and zoom-in (right)

26
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Both graphs clearly demonstrate that, whereas a price of DSF. The difference between them is defined as the
floor already exists in the no-flex scenario (due to large DSF adequacy benefit. The following assumptions have
amounts of electrolysers and front-of-meter storage), the been made:
DSF scenario further reduces price volatility, rendering a
relatively flat price duration curve for 2030. Additional generation capacity does not come
from carbon-free technologies, since there are more
Adequacy considerations cost-effective ways for decarbonising the system than
constructing carbon-free generators that run less than
The analysis above shows that the no-DSF scenario does 100 hours per year. Therefore, the price of gas peaking
not maintain security of supply, showing a loss load of plants is considered.
2,054 GWh. In terms of capacity, the no-DSF scenario
lacks up to 60 GW of generation capacity during the For DSF, the costs of industrial DSF are considered,
highest peak in residual load. In the DSF scenario, the as these are likely to play a dominant role in scarcity
system fulfils the lack of generation capacity with load situations. This is limited to the (annualised) enablement
shifting and load curtailment. costs of DSF. Typically, industrial customers also require
annual (capacity) payments for participating in services
While a full adequacy analysis was not the scope of this with low activation frequencies. Because these payments
study13, this analysis gives an indication of the minimum are direct benefits for consumers, they are not considered
gap (60 GW) that would need to be covered by additional as additional costs to consumers.
peak generation. Since the DNV model only considers
marginal costs, it does not quantify the capital expenditure The table below gives a rough indication of the capital
(CAPEX) benefits related to generation adequacy, i.e. investment needed for both options. The gas peaker
avoiding the construction of (in this case approximately option is substantially more expensive than the investment
60 GW) additional capacity. Therefore, DNV calculated needed for installing 60 GW of DSF. Therefore, DNV
an indicative adequacy benefit of DSF by comparing 1) concludes that, roughly, the adequacy benefit of DSF
the investment required for installing 60 GW of peak in 2030 is €2.7 billion.
generation capacity; and 2) the costs of enabling 60 GW

Total cost
Year 2030 – EU 27 Cost [€/MW/year] [million €]

Gas peaker CAPEX (annualised) 45,500 (LAZARD, 2020) 2,730

DSF CAPEX 120 (European Commission DG Energy, 2016) 7.2

13. An adequacy study needs stochastic modelling to draw statistically valid conclusions. The modelling used for this study was deterministic, and therefore not sufficient.
27
 4.2. System balancing
DNV’s market model calculates day-ahead (DA) spot prices, and does not consider balancing energy and associated
costs. Therefore DNV has quantified the savings, that DSF could potentially bring to balancing markets by 2030,
following a simplified methodology presented in Appendix A – section 1.2. The quantification considers the difference
between energy balancing costs of the DSF and no-DSF scenarios in 2030. Given the relatively small size of balancing
markets compared to wholesale markets, DNV performed the calculation under four main assumptions:

1. By 2030, balancing energy will be procured at European level, not country level;

2. There is sufficient interconnection capacity so all balancing resources at European level are available to all Member
States, therefore assuming that current restrictions regarding excessive reliance on importations of balancing
services no longer apply;

3. Balancing capacity costs and balancing energy utilisation will remain at current levels; and

4. The balancing energy costs are determined by the marginal costs of technologies technically capable of providing
the different balancing services.

Under these assumptions, DNV first identified DSF and no-DSF technologies that could technically provide the
different balancing services and their associated marginal costs. DNV then built a merit order for each balancing
service for each scenario (DSF and no-DSF). Subsequently, based on the required balancing capacity, DNV identified
the technologies that would cover the balancing needs. Finally, the costs for balancing energy were calculated for
a full year based on the marginal costs of the technologies ‘in the money’ as well as the savings, i.e. the difference
in cost between scenarios.

DNV’s calculations showed that the participation of DSF technologies in European balancing markets in 2030 could
save, in total for EU 27, between €262 million and €690 million. In relative terms, this translates into a balancing
energy cost saving of between 43% and 66%. The wide range of savings is due to the uncertainties related to the
different balancing capacity needs for all EU 27 members. The different capacity needs correspond to different price
levels in the balancing merit order, which in turn translate into a wider saving range.

It is worth noting that one of the limitations of the approach is that the quantification does not consider opportunity
costs for participating in other markets. A more elaborate overview of the approach is presented in Appendix A –
Section 1.2.2.

28
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

4.3. Grid infrastructure

As previously described in section 2, DNV’s market model does not include distribution grid restrictions or local
flexibility markets. Therefore, DNV has created a simplified methodology, based on the limited data available, to
quantify the required investments in distribution grids between 2023 and 2030 and the infrastructural benefits of
DSF. A detailed description of the methodology is presented in Appendix A – Section 1.1.
The methodology is built under these main assumptions:

1. All available DSF flexible power behaves in a grid-friendly manner. This means that flexibility responds to network
needs either through (collective) self-consumption, grid tariff optimisation or provision of flexibility services to
distribution system operators.

2. Distribution grid investments are mainly driven by final electricity demand, adoption of EVs and RES capacity.

3. The potential savings are proportional to the ratio of DSF available capacity and peak load.

In summary, this methodology consists of both a top- power and the investment savings, DNV estimates that
down and bottom-up phase. In the former, the required DSF can enable savings between €77.6 billion and €203.6
distribution grid investments at EU 27 level are calculated billion (i.e. between -27% and -80%) at EU 27 level
and distributed across Member States according to their between 2023 and 2030. Assuming that the annual
share of electric load, electric vehicles, and RES capacity investments are constant, the annual savings in 2030 are
connected to the distribution grid. In the latter, based on estimated between €11.1 billion and €29.1 billion. The
available DSF capacity relative to peak load, savings in large range of potential savings is due to the top-down
infrastructural investments are estimated per Member approach taken to quantify the grid investment needs
State and then aggregated at EU 27 level. Investments per Member State. This is due to the uncertainties on
required for the modernisation, digitalisation, automation, the extent to which each country will contribute to the
and resilience of the distribution grids are not accounted total estimated investments, and hence to the benefits
for. enabled by available DSF.

According to (Eurelectric, 2021), investment needs The quantification approach proposed by DNV focuses
in distribution grids are mostly driven by the final on the investments required in low-voltage and medium-
electricity demand, the number of electric vehicles, voltage distribution grids only and does not include
and the renewable capacity connected to the distribution high-voltage grid or transmission grids. This means that
grid.14 Based on the input data, calculated DSF capacity potential savings on transmission grid reinforcements
and the investment breakdown per driver derived from or DSF contribution to lowering redispatch costs have
(Eurelectric, 2021), DNV estimates that the EU 27 required not been quantified. Additionally, it does not consider
investments in distribution grids to integrate new loads differences across countries regarding the current status
and RES capacity are between €253.1 billion and €282.5 of development of the distribution grid and voltage level
billion between 2023 and 2030. in the distribution grid. A more elaborate overview of the
approach limitations is presented in Appendix A – section
The results of the grid simulations in (E-Bridge, 2019) 1.3.3. Overall, DNV acknowledges the lack of studies on
show that the grid-friendly use of DSF capacity can reduce the quantification of infrastructural benefits of DSF. DNV
by 76.9% the required investments in low-voltage and interprets this as a signal for relevant stakeholders to
medium-voltage distribution grid expansion in Germany further investigate the topic.
by 2035. Assuming a linear relation between DSF available

14. These grid investment needs refer particularly to the needs raised by higher electrification and connection of renewables.
29
 QUANTIFICATION OF
5. CONSUMER BENEFITS
Here, the DSF benefits at system level calculated in the previous section are translated into benefits to the consumer.
Direct benefits end up at residential, commercial and industrial consumers with flexible loads, for example, EV
owners who exploit the flexibility of their EVs face lower energy costs when EV charging is exposed to dynamic
prices. Additionally, all consumers will gain indirect benefits, due to the overall effect on market prices.

5.1. Direct benefits

The direct benefits are those that the consumer sees Table 5.1 includes a summary of direct benefits to
directly reflected in their energy bill15, as either a saving consumer.
or a revenue (e.g. batteries). Thus, these benefits can
be split per type of technology in 2030, which gives an The highest savings are achieved on space electric
average indication of the savings per energy unit. DNV heating with a potential cost reduction of €71 billion in
calculated the savings/revenues depending on the type of 2030 for EU 27. These savings could be even greater when
DSF technology, excluding capital investments for enabling taking energy-efficiency considerations into account. Some
or acquiring the different DSF technologies: studies indicate that residential flexibility activations also
accomplish a reduction of the overall load. For instance, an
Smart charging and space electric heating: The savings RTE study (RTE, 2016) suggests that energy savings up to
are calculated as the difference between the cost to 50% could be achieved. However, DNV has not modelled
source their required amount of energy in the DSF and these aspects due to lack of evidence applicable to all
no-DSF scenarios in 2030. Member States throughout a full year.

Battery storage and V2G: The savings are calculated An average EV owner could save up to 0.07 €/kWh,
as the difference between battery/V2G profits for both which adds up to a saving of 176 €/year on their energy
scenarios in 2030. The profits are defined as the difference bill in 2030. The average includes both smart charging
in revenue generated by energy infeed and the costs and V2G savings.
(charging and variable maintenance and operation costs).
It can also be observed that behind-the-meter revenues
Industrial electric heating: The savings are calculated are significantly lower than the rest. Partly, this is due to
as the difference between the energy costs of both efficiency cost considerations.
scenarios. The difference in energy costs is caused by the
load curtailment in the DSF scenario and the different Finally, industrial DSR figures show the savings that can
energy prices. be achieved through (market-based) load curtailment
only, i.e. the reduction in total energy payment due to
Industrial DSR: The savings are calculated as the different energy prices is not included in the calculations.
reduction in energy costs due to curtailment minus the
cost incurred due to curtailment.

15. This is assuming that these customers are exposed to dynamic prices that reflect wholesale prices. Other remuneration models are also conceivable, e.g. fixed energy prices with a
30 separate remuneration for providing flexibility.
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Table 5.1 – Direct consumer benefits in the DSF- versus the no DSF-scenario

Savings and % Relative Average

Year 2030 – EU 27 revenues to no-DSF saving/revenue
(million €) costs (%) per kWh (€/kWh)

Smart charging + V2G 9,936 48% 0.07

Space electric heating 71,234 64% 0.16

Battery storage BTM 32 100% 0.08

Industrial electric heating 5,166 51% 0.05

Industrial DSR (curtailment) 232 100% 0.01

5.2. Indirect benefits

The changes in system costs and energy prices will have avoids the use of more expensive generation. Following
an impact on all final consumers. The indirect potential the energy prices, and with a lower order of magnitude,
benefits that DSF provides to the consumer are: are infrastructure savings per year, representing €27.8 to
€65 per capita in year 2030. The savings with the lowest
The overall reduction of energy prices (i.e. the reduction impact are balancing services, at €0.7 to €1.6 per capita
of cost to serve consumer’s load) in year 2030; these savings are relatively low because the
The reduction of generation capacity costs size of the market is significantly smaller compared with
The decrease in investment for grid infrastructure wholesale. Finally, one of the benefits for all consumers
The decrease of system balancing costs is the lowering of carbon emissions by 83.8 kg per capita,
The decrease of carbon emissions which would fulfil the 55% reduction by 2030.

Based on the results from the previous sections, Table Whereas end consumers benefit from reduced electricity
5.2 summarises the indirect impacts to all consumers, market prices, generators will substantially lose margin
in absolute terms and per capita. The results indicate on their transactions. This effect is the strongest for
that the most significant saving is due to the reduction fossil, nuclear and biomass power plants; the impact for
of energy costs, with an average reduction of €673.5 in renewables is substantially less as these can also benefit
energy cost per capita in year 2030. As previously detailed from additional load during times of high renewable
in section 4.1, this is due to the effect of DSF in peak generation. An overview of generator margins and further
shaving and reduction of renewable curtailment, which considerations in the implications are presented in section 4.1.

Table 5.2 – Indirect consumer benefits

Potential
Potential % Relative
Year 2030 – EU 27 savings
savings to no-DSF per capita16

Cost to serve load €301.5 billion -48% €673.5

Adequacy €2.7 billion -100% €6.0

Balancing €0.3–0.7 billion [-66%, -43%] €0.7–1.6

Infrastructure17 €11.1–29.1 billion [-80%, -27%] €27.8–65

Emissions 37.5 Mt -8% 83.8 kg

16. Savings per capita are calculated for 447.7 million inhabitants in EU 27 (eurostat, 2020).
17. It should be noted that the reference for the calculation of infrastructure benefits is not the ‘no DSF’ scenario. This is due to the methodology applied to this
calculation. The savings and relative calculation are based on the investments needed as predicted today in a business-as-usual situation.
31
 While the above categories are all potential benefits of DSF, they are not stackable and therefore the benefits cannot be simply
summed up to a total. The next section provides further insight on how these results can be interpreted.

5.3. Stackability considerations

In its methodology, DNV has assessed the value of DSF in the different segments separately, – i.e. wholesale, generation adequacy,
balancing and infrastructure. Analysis shows that the potential value is significant, especially for the costs to serve load.
This study assumes that all DSF technologies have access to these segments and will be activated based on a market-based
mechanism – provided that they meet the technical requirements (mainly balancing) and geographical requirements (mainly
infrastructure).

For estimating the total value of DSF, the three separate outcomes cannot simply be added. Although ‘value-stacking’ (participating
in different markets at the same time) of DSF in general is allowed, there are clear physical limitations to this concept, for example:

If a flexible EV charger is charging at maximum capacity due to low wholesale prices, it cannot increase its load to provide
balancing power (it can only provide balancing power in one direction in this case).

More delicate is the combination of wholesale and infrastructure. The deployment of DSF can improve both segments at the
same time but can also create conflicts. Considering the same EV charging example:

Optimising EV charging against wholesale prices will reduce EV charging load during late-afternoon peak hours, reducing the
stress on grid infrastructure in (urban) load centres;

Optimising EV charging against wholesale prices will increase EV charging during periods around noon with high solar-PV
power production, reducing the need for (market) PV curtailment, but increasing the stress on grid infrastructure in solar-PV
dominated (rural) areas.

The DNV model has not taken any limitations on the distribution grid into account. This means that when DSF is deployed to
reduce infrastructure costs, this will lower the benefits for the cost to serve load. Since the potential benefits on infrastructure
are significant, albeit lower than benefits for the cost to serve load, there is a compelling reason to look for the middle ground
of market and infrastructure optimisation. Using a proper market design, where all DSF not only is exposed to market prices as
well as cost-reflective grid fees and DSO flexibility services, can ensure that the benefits for the customer are optimal, and will
exceed the value of benefits to the cost to serve load only.

Generation adequacy benefits have been calculated outside the wholesale model by considering CAPEX costs of generation
assets that would be needed to solve the generation deficiency of the no-DSF scenario. Including these assets for the no-DSF
scenario within the model would lower the calculated wholesale benefits.

32
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

KEY TAKE-AWAYS AND

6. RECOMMENDATIONS

In 2030, the activation of 397 TWh upward DSF and 340.5 that the highest savings are achieved in space heating, with
TWh of downward DSF enables Fit for 55 and REPowerEU a 64% cost reduction, followed by EV charging with a 48%
strategy by 2030 while achieving a cost reduction on: cost reduction (€0.07 per kWh saving). The lowest savings
are achieved by DSF technologies with higher variable
Wholesale and adequacy operation and maintenance costs, such as industrial DSR
and, to a lesser degree, batteries. These benefits are also
- The cost to serve load is up to around €301.5 billion reflected in the indirect benefits; therefore, they should
(48%) less; not be double-counted.

- The emissions are up to 37.5 Mt (8%) less; Indirect benefits: The results indicate that the most
significant saving is due to the reduction of electricity
- Costs to generate are up to €4.6 billion (5%) less; costs, reducing the costs to consumers by up to €673.5 per
person in the year 2030. This is due to the effect of DSF
- Renewable energy curtailment is up to 15.5 TWh in peak shaving and reduction of renewable curtailment,
(61%) less; and which avoids the use of more expensive generation.
Following the electricity prices, and with a lower order
- The costs saved on installed generation capacity are of magnitude, are infrastructure savings of between
up to €2.7 billion less. €27.8 and €65 in cost reduction per capita in a year in
2030. The savings that would make the least impact are
Balancing energy costs are estimated to be between those in balancing costs (€0.7 to €1.6 per capita in 2030);
€262 million and €690 million less (43% and 66%) these are relatively low because the size of the balancing
market is significantly smaller than wholesale trade. Finally,
Grid infrastructure savings in low-voltage and medium- one of the benefits for all consumers is the reduction of
voltage are between €77.6 billion and €203.6 billion (i.e. carbon emissions. The DSF scenario would fulfil the 55%
between-27% and-80%) at EU 27 level between 2023 and reduction by 2030, while the no-DSF would not.
2030. The annual savings in 2030 are estimated to be
between €11.1 billion and €29.1 billion. These savings DNV calculated the savings that full DSF enablement (~160
are based on the hypothesis that grid-driven optimisation GW in capacity across EU 27) by 2030 would achieve
is implemented, which may not always be the case in the against a scenario in which DSF is not available. The lack
above-mentioned market-driven optimization cases. The of data on the utilised DSF today and DSF utilisation
potential savings on transmission grids due to avoided prospects towards 2030 makes it impossible to quantify
reinforcement and re-dispatch costs are not quantified. the counterfactual system costs. Therefore, DNV modelled
a 2030 scenario in which DSF technologies are not flexible.
This translates into direct and indirect consumer benefits. Realising that demand-side technologies are to some
The direct benefits are achieved by a direct saving due to extent – and in some Member States more than others
the shifting or curtailment of the load, whereas indirect – already actively participating in the electricity market,
benefits are achieved due to the impact of DSF activation DNV acknowledges that the reference scenario is not
in the entire system and its energy costs. a realistic scenario. Therefore, all results, should be
interpreted as the total potential benefit of DSF. If, by
Direct benefits: The users of DSF technologies can save 2030, certain barriers still exist to DSF fully participating
between €0.01 and €0.16 per kWh. The results indicate in the electricity market, a certain share of this potential

33
 will not be achieved. The results provide an order-of-
magnitude value of this missed opportunity (albeit an
upper bound).

The study presents several limitations, partly due to data

unavailability at European level. Moreover, the results
raised some questions that are worth analysing further.
Therefore, DNV recommends to:

Improve transparency on the current deployment of

DSF in wholesale and balancing markets throughout EU
27 by introducing market monitoring in a structured and
harmonised manner on DSF availability and deployment;

Assess and mitigate any regulatory, economic (e.g. grid

tariff methodology), and/or business (case) challenges
that may cause bottlenecks to further DSF deployment;

Develop a common methodology to quantify the

potential infrastructure savings resulting from a full
deployment of DSF per Member State;

Analyse which holistic market design (including grid

fee methodology and DSO flexibility services) can achieve
the optimal DSF deployment results for the combination
of both markets and infrastructure;

Investigate energy-efficiency potential resulting from

DSF deployment; and

Investigate the impact of more large-scale DSF

deployment on the investment climate for renewables
generators.

34
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

APPENDIX A – Detailed
Methodology Description

1.1. Quantification of consumption or to reduce grid fees. This is not explicitly

considered, because
wholesale and adequacy
benefits - the grid benefits are determined separately (outside
the model), this model focuses on the market benefits.
1.1.1. QUANTIFICATION APPROACH
- In practice, the behaviour triggered by in-building
DNV has used its European Market Model which is a optimisation, is very similar to the behaviour triggered
fundamental market model that simulates the day- by market optimisation. For example, a home battery
ahead spot price by optimising the unit commitment will typically charge around noon on a summer day
and economic dispatch of electricity generation. The whether there is local PV (in-building optimisation)
simulations are performed on an hourly time-resolution or not (market optimisation).
containing a detailed representation of generation,
commodity prices and demand for all bidding zones Time shifters (e.g. electric heating, EV charging and
in EU 27 Member States. The description of the model batteries) do not have marginal costs, they are simply
inputs is described in section 2. optimised against wholesale prices within time constraints.
Battery and V2G need to take efficiency losses into account,
Among other parameters, the market model outputs the so there should be sufficient spread between charging
hourly demand and generation mix, the hourly day-ahead and discharging prices to cover these costs.
energy prices, and the total system emissions.
Wholesale
DNV has used the outputs of the model to quantify the
wholesale benefits. To do that, DNV has created two a Using the output of the model, DNV calculates the
reference scenario to compare the 2030 results (“DSF wholesale benefits by comparing the DSF and no-DSF
scenario”) – with full DSF potential – against. As explained scenario:
in section 2.4, DNV modelled a 2030 scenario in which
the demand side is not flexible to serve as a reference Total carbon emissions in 2030: This is a direct output
scenario, albeit not realistic. of the model and it’s based on the carbon-based fuel
and biomass used by the dispatched generators. The
In the next section (1.1.2), there is a detailed explanation carbon capture and storage impact is also considered in
on the approach taken to model the selected DSF the results. The resulting emissions are then evaluated
technologies in the DSF and no-DSF scenarios. The main against the targeted 2030 power system emissions to
underlying assumptions to our modelling are: fulfil the 55% emission reduction objective. This power
sector emission target is based on DNV’s energy transition
DNV considers the full theoretical potential of DSF outlook model and it’s set at 410 eq. CO2 million tonnes.
of the selected technologies, although in practice some
consumers may not be interested to participate. Total curtailment volume in 2030: This is calculated by
subtracting the realised solar and wind generation to the
Some flexibility may not be available to the market, as it available generation in a given hour, for all the hours of the
is used for ‘in-building’ optimisation, e.g. to increase self- year. The curtailment can be due to economic reasons or

35
 interconnector capacity limitations. The model does not 1.1.2. DEMAND AND DSF MODELLING
consider curtailment due to distribution and transmission APPROACH AND INPUT PARAMETERS
constraints within the bidding zone.
In this section, the modelling of the different technologies
Total generator costs: This is a direct output from is explained in detail, for both the DSF and no-DSF
the model. These are the costs that generators/storage scenarios.
incur to cover the system demand. These include fuel
costs, variable operation and maintenance (VOM) costs, 1.1.2.1. Industrial DSR
start and shutdown costs, emissions costs and penalties. Industrial DSR is modelled as a generator whose generation
Penalties apply when a generator violates a constraint, corresponds to an industrial load (partially) shutting down
e.g. a minimum generation requirement. its consumption. Industrial DSR is defined per country and
distinguished in five types of industrial load. Each type is
Loss load (unserved load): This is calculated as the defined by the following parameters:
sum of the hourly demand that is not met in 2030.
Installed capacity [MW]: max generation, i.e., max
Cost to serve load: This is calculated for all loads in a curtailable load. It is defined as the total industrial DSR
year and is defined as the price paid by the load times capacity multiplied by a percentage that varies per type.
the demand. The price paid by the load is the volume-
weighted price paid for all energy purchases in 2030. Variable O&M [€/MWh]: cost of shutting down the
industrial load
Adequacy
Max Up Time [h]: maximum number of consecutive
To have a rough estimation of DSF adequacy benefits, hours the industrial load can be curtailed
DNV also relies on the market model simulations results. While the total industrial DSR capacity varies per country,
The approach is as follows: the distribution across load types, the variable O&M, and
the max up time are the same, as shown in Table 1.1.
1. Calculate the demand power that was not served
by generation in the no-DSF scenario.

2. Calculate the CAPEX needed to install a

generation plant that can cover the generation
power deficiency.

3. Calculate the CAPEX needed to enable the same

power, as in step 2, of industrial DSR.

4. The rough estimate (lower bound) of DSF

benefit to generation adequacy is calculated as the
difference between step 3 and step 2 results.

36
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Table 1.1 – Industrial DSR parameters per load type

Percentage Variable Max Up

Type of DSR of total DSR capacity O&M [€/MWh] Time [h]

Category 1 10% 1000 1

Category 2 35% 1097 2

Category 3 9% 323 4

Category 4 30% 500 8

Category 5 16% 207 -

The values for the abovementioned parameters are derived from several sources, which are listed in Table 1.2
– Industrial DSR data sources. Countries not listed in Table 1.2 are expected to have no industrial DSR installed
capacity in 2030.

Table 1.2 – Industrial DSR data sources

Percentage Variable Max Up

Country Total installed capacity Note
per type O&M Time

AT (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

BE (Elia, 2021) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

DE (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

EE (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

ES data are not reported in

ES (ENTSO-E, 2020) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019) (ENTSOG and ENTSO-E, 2022)

FI (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

FR (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

HR (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

IE (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

IT (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

LT (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

LV (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

NL (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

SE (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

SI (ENTSOG and ENTSO-E, 2022) (Elia, 2019) (ENTSO-E, 2021) (Elia, 2019)

37
 The Variable O&M is derived from (ENTSO-E, 2021) by calculating the weighted average price across countries per
DSR type according to the max up time value.

No industrial DSR units are modelled in the no-DSF scenario.

1.1.2.2. BESS
Batteries are distinguished in behind-the-meter and grid- BESS charge and discharge depending on the power price
connected BESS. Both types are defined by the following (see illustrative example in Figure 4), while satisfying
parameters: some constraints, namely, maximum charging/discharging
power, maximum storage capacity, minimum SoC, and
Units: number of units (following properties are defined maximum cycles per day. For the battery to be activated,
per unit). the power price needs to show a sufficiently large spread
Max Power [MW]: charging/discharging max power in one day look-ahead to overcome the cost of energy
capacity per unit. due to efficiency losses.
Capacity [MWh]: capacity of each unit. A battery storage
duration of 3 hours is assumed, which is comparable to The values of the abovementioned parameters are mostly
Li-ion technology. based on expert knowledge and are summarised in Table .
Min SoC [%]: minimum state of charge of a unit. The number of units per BESS type per country are derived
Initial SoC [%]: initial state of charge of a unit. from multiple sources, which are listed in Table 1.4.
Charge/discharge efficiency [%].
Max Cycles Day: number of cycles allowed each day.

Table 1.3 – BESS techno-economic parameters

Capacity Min SoC Initial SoC Charge/discharge Max cycles

Max Power [MW] [MWh] [%] [%] efficiency [%] day

100 300 20 50 85 2

Table 1.4 – BESS data sources

Parameter Source

Total number of batteries per country (ENTSO-E, 2020), (DNV, 2021)

Split between behind-the-meter and grid-connected (EASE)

1.1.2.3. Smart charging Max load – maximum hourly (charging) capacity

EVs are modelled using the following parameters: The maximum hourly (charging) capacity is calculated
based on the maximum power capacity per country and
Max load [MW]: maximum hourly (charging) capacity. an availability profile. The maximum power capacity per
Max Daily Consumption [GWh]: maximum daily load country is calculated based on the numbers of charging
Min Daily Consumption [GWh]: minimum daily load points per country times the maximum power capacity
per car. The model assumes 10 kW/car.
EV charging load is optimised by the Market model.
However, it is subject to a daily load equality constraint that The availability profile represents the share of charging
cannot be violated, i.e., max energy day = min energy day. points occupied by a car per hour.

38
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Max daily consumption and Min daily consumption – maximum/minimum daily load

The minimum and maximum daily load is calculated based on the annual EV consumption per country divided by 365.
The annual EV consumption per country is calculated based on the number of EVs times the annual consumption
per car, which is assumed equal to 2.5 MWh/year.

An overview of the input parameters is given in Table 1.5.

Table 1.5 – EV technical parameters

Max energy capacity Annual consumption

Max Power [MW] [kWh/car] [MWh/year/car]

10 61.5 2.5

In the no-DSF scenario, EVs follow an hourly charging profile which they cannot deviate from.

The abovementioned parameters and availability profiles are derived from multiple sources which are listed in
Table 1.6.

Table 1.6 – Smart charging data sources

Parameter Source Note

Number of EVs per country (DNV, 2021) and industry insights BEV and PHEV data

Number of CPs per country (DNV, 2021) and industry insights Home and workspace data

Availability profile (ElaadNL, 2020), (UK Government, 2022),

(California Institute of Technology, 2021)

Charging profile (no-DSF) (ElaadNL, 2020), (UK Government, 2022),

(California Institute of Technology, 2021)

1.1.2.4. V2G Capacity

It is assumed, based on expert knowledge, that 30% of the
EV charging infrastructure provides V2G possibility. V2G It corresponds to the storage capacity available for V2G. It
is modelled as batteries with the following parameters: is calculated based on the number of EVs per country times
the max energy capacity per car, and on the assumption
Capacity [MWh] that 30% of the grid connected EVs provides V2G capacity.
Max power [MW]: power at full discharge Max power and max load – power at full discharge/charge
Max load [MW]: power at full charge It corresponds to the power capacity available for V2G.
Initial SoC [%] It is calculated based on the EV max load (see Section
Charge/discharge efficiency [%] 1.1.2.3) and on the assumption that 30% of the EVs
Max Cycles Day: number of cycles allowed each day. provides V2G capacity.

An overview of the input parameter is given in Table 1.7. Data sources are the same used for smart charging (see
Table 1.6).

39
 Table 1.7 – V2G techno-economic parameters

Charge/discharge Max cycles

Initial SoC [%] efficiency [%] day

50 93 2

V2G is not modelled in the no-DSF scenario.

The abovementioned parameters are derived from multiple sources which are listed in Table 1.8.

Table 1.8 – EV data sources

Parameter Source

% V2G charging infrastructure (European Commission, Directorate-General for Energy, 2022)

1.1.2.5. Residential electric heating Min and Max load 12 h – maximum/minimum

Residential electric heating represents residential and consumption in 12-hour period
commercial space heating electricity demand. It is
considered as a shiftable load, and it is modelled using The minimum and maximum daily load is cal
the following parameters: culated based on the annual residential electric heating
consumption per country, distributed among the days
Max Load [MW]: maximum hourly consumption. based on a daily space heating demand profile. This
Min Load 12 h [GWh]: maximum total space heating daily consumption is divided in two periods of 12 hours
consumption in 12-hour day/night period. each, i.e. half of the total daily consumption is required
Max Load 12 [GWh]: minimum total space heating in each 12-hour period.
consumption in 12-hour day/night period.
The 12-hour periods are defined as:
Residential electric heating load is optimised in the Market
model. However, it is subject to a 12-hour load equality Day period: from 9 a.m. to 9 p.m.
constraint that cannot be violated, i.e., min load 12 h Night period: from 9 p.m. to 9 a.m.
= max load 12 h. Therefore, it can provide flexibility by
shifting the consumptions while respecting the 12 hours The minimum and maximum 12-hour load is defined
load requirement. with the same value. Therefore, the consumption can
be shifted across a 12-hour timeframe, but the required
Max load – maximum hourly consumption consumption (minimum and maximum) must be fulfilled
in those 12 hours.
Max load corresponds to the maximum possible
consumption from residential space electric heating in In the no-DSF scenario, residential electric heating follows
every hour, per country. The maximum load is the same an hourly consumption profile per country, from which
for every hour. there cannot be deviations. Hence, demand cannot be
shifted across time.

The abovementioned consumption values and daily

profiles are derived from multiple sources which are
listed in Table 1.9.

40
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Table 1.9 – Residential electric heating data sources

Parameter Source Note

Annual electricity demand (ENTSOG and ENTSO-E, 2022) Residential Space heating and
Tertiary space heating data. Scenario
Distributed Energy.

Daily and hourly heat (Ruhnau, O., Hirth, L. & Praktiknjo, A.) Single-family house, multi-family
demand profiles house, and commercial building data.

1.1.2.6. Industrial electric heating 1.1.2.7. Industrial heating – CHP

Industrial electric heating is considered as a curtailable Industrial heating from CHP units is represented by CHP
load, and it is modelled using the following parameters: generators with hourly generation requirements. The
units can deviate from the generation requirements by
Max Load [MW]: maximum hourly consumption. incurring certain penalty. These units are modelled as
Bid Quantity [MW]: demand bid. standard generators, with several additional parameters:
Bid Price [€/MWh]: price of the demand bid.
Min Energy Hour [MWh]: minimum hourly generation.
Industrial electric heating load can reduce part of its Min Energy Penalty [€/MWh]: penalty for not reaching
consumption when prices reach a certain activation price. the minimum hourly generation.

Max load – maximum hourly consumption Min Energy Hour – minimum hourly generation.

Max load corresponds to the maximum possible The minimum generation per hour is defined depending
consumption from industrial electric heating in every on the hour of the day:
hour, per country. The maximum load is the same for
every hour. Peak hours: from 5 a.m. to 8 p.m. the unit should
generate at 100% of its capacity.
Bid quantity – demand bid
Off-peak hours: from 8 p.m. to 5 a.m. the unit should
Bid quantity represents the hourly consumption from generate at minimum 65% of its capacity.
industrial electric heating, and is defined with the
same values as Max Load. Therefore, bid quantity is the Min Energy Penalty – penalty for not reaching the
hourly consumption and also the maximum possible minimum hourly generation
load shedding.
If the generation does not reach the required level defined
Bid price – price of the demand bid by Min Energy Hour, then a penalty is incurred. The value
of this penalty is defined as 1000 €/MWh.
Bid prices represents the activation price for industrial
electric heating curtailment. The load will be curtailed In the no-DSF scenario, industrial heating CHPs cannot
if the prices reach this specific bid price. The bid price is deviate from the generation requirements described
defined as 500 €/MWh. above. Deviations by either not reaching or surpassing
the requirement are not allowed.
In the no-DSF scenario, industrial electric heating cannot
curtail the consumption. The values of the abovementioned parameters are mostly
based on DNVs internal data.
The values of the abovementioned parameters are mostly
based on expert knowledge and DNVs internal data.

41
 1.1.2.8. District heating – CHP 1.1.2.9. Electrolysers
District heating residential CHP units are represented Electrolyser consumption is considered as a curtailable
by CHP generators with daily generation requirements load, and it is modelled using the following parameters:
which are linked to the daily heat demand. CHP units are
defined as part of a heating area for which they must fulfil Max Load [MW]: maximum hourly consumption.
certain level of generation per day, which represents the Max Yearly Consumption [GWh]
required heat consumption. These units are modelled as
standard generators, with several additional parameters: Max Load [MW]: maximum hourly consumption.

Min daily generation [GWh]: minimum daily generation. Bid Quantity [MW]: demand bid.
Min daily generation penalty [€/GWh]: penalty for not Bid Price [€/MWh]: price of the demand bid.
reaching the minimum daily generation.
Electrolyser load can reduce part of its consumption
Min daily generation – minimum daily generation when prices reach a certain activation price. Moreover,
the annual consumption defined needs to be fulfilled.
The minimum daily generation of each CHP unit is based However, the hours when this consumption occurs are
on the minimum annual generation requirement for that optimised by the Market model, based on the power
heating area. This requirement is defined as the installed prices and system requirements.
capacity of each CHP belonging to that heating area,
multiplied by the average CHP full load hours. DNV has Maximum yearly consumption
considered 3000 full load hours.
The maximum annual consumption is determined for every
The annual requirement is translated into a daily constraint, country, and it is based on the REPowerEU Communication
by distributing the annual minimum generation among that presents a target of 10 million tons of renewable
the days of the year, based on a daily heating demand hydrogen production in Europe.
profile per country.
Max load – maximum hourly consumption
Min daily generation penalty – penalty for not reaching
the minimum daily generation Max load corresponds to the maximum possible
consumption from electrolysers in every hour, per country.
The CHP units can deviate from the minimum daily The maximum load is the same for every hour. This value
generation by incurring a penalty. This penalty represents is based on the EU27 estimated Max yearly consumption
the replacement of the CHP generation by a gas boiler. and the electrolyser utilization factor. The electrolyser
The value consider for this penalty is 41914.6 €/GWh. capacity for each county is distributed based on the ratio
traditional load country/traditional load EU27.
In the no-DSF scenario, district heating CHP cannot
deviate from the generation requirements described Maximum yearly consumption
above. Deviations by either not reaching or surpassing
the requirement are not allowed. The maximum annual consumption is determined for every
country, and it is based on the REPowerEU Communication
The values of the abovementioned parameters are mostly that presents a target of 10 million tons of renewable
based on expert knowledge and DNVs internal data. hydrogen production in Europe.

Bid quantity – demand bid

Bid quantity represents the hourly consumption from

electrolysers, and it is defined with the same values as Max
Load. Therefore, bid quantity is the hourly consumption
and also the maximum possible load shedding.
Bid price – price of the demand bid

42
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Bid prices represents the activation price for electrolyser load curtailment. The load will be curtailed if the prices reach
this specific bid price. The bid price is defined as 86.2 €/MWh, based on the cost of producing hydrogen. (RTE, 2016)

In the no-DSF scenario, electrolyser load is defined, additionally to the parameters mentioned above, with a Min
Energy Year requirement. This parameter is defined with the same data as the maximum energy year property.
Hence, the annual electrolyser load is fixed and cannot be curtailed, but load shifting can still occur.

The abovementioned annual targets are derived from the following sources:

Table 1.10 – Electrolysers data sources

Parameter Source

Annual maximum/minimum consumption (European Commission, 2022)

DNV, (Planbureau voor de Leefomgeving, 2020),

Hydrogen pricea (Thomas D. (Hydrogenics), 2016)

1.1.3. LIMITATIONS OF THE APPROACH Potential flexibility from public charging points not
included, based on the assumption that most cars
DNV acknowledges that the proposed approach for the connected to public chargers will not be able to provide
quantification of market and adequacy benefits shows a lot of flexibility as drivers will want to ‘move on’.
some limitations which are briefly presented below:
Apart from EVs, other means of electric transport such
No-DSF scenario and modelled flexible technologies: as electric buses or trucks are not modelled. This is due
The no-DSF scenario does not provide a realistic to data unavailability.
counterfactual, but rather a reference scenario that
estimates the full available potential. At the same time, Investment costs are not considered for either scenario.
due to data unavailability, DNV did not include a few
relevant DSF technologies in their modelling. Therefore, The category Residential electric heating represents
the savings are the upper bound of savings considering space heating demand and includes the electricity demand
the modelled, but not for ALL potential DSF capacity. of heat pump technologies for residential and commercial
buildings. Other technologies are not included in this
Industrial DSR marginal costs: Despite more detailed category.
insight for the Netherlands in terms of industrial process
and cost details, DNV decided to exclude this from the DNV has taken a conservative approach and considered
model and instead work with the assumptions provided residential electric heating as a shiftable load. An RTE
the above sources, that are less process-specific. report (RTE, 2016) suggests that up to 50% of residential
Predominantly because a detailed country-by-country electric heating can be saved when DSF is activated,
analysis regarding (details of) industrial processes goes i.e. residential electric heating is not only shiftable
beyond the scope of this research. The used figures are but curtailable. The model has not incorporated the
therefore assumed to be more conservative. For example: quantification insights of this report since it is based
In the Netherlands, our previous studies indicated that on a different technology (Joule effect heating), limited
there is relatively cheap DSR capacity from chlorine period of the year and limited geography. However, DNV
production. Dutch chlorine production is only a few % acknowledges that residential heating flexibility could
of what’s produced in Europe, but we DNV did not find provide additional savings, next to the savings that are
any evidence of chlorine production in other countries provided by load shifting according to the results.
being used to provide DSR (at an equally low cost).

43
 Due to the high electrolyser installed capacity, the electrolyser bidding price has a high impact on system prices.
Although the bidding price highly depends on the characteristics from individual electrolyser plants, the model
assumes a constant bidding price throughout Europe.

The flexibility that could be provided by cooling technologies has not been considered in this analysis. This was
due to limited data availability at European level.

The renewable curtailment presented in this report refers to the curtailment due to higher available generation
than load and considers the possible congestion of the interconnectors. However, congestion in national distribution
and transmission networks is not considered.

1.2. Quantification of system balancing benefits

1.2.1. QUANTIFICATION APPROACH

DNV calculated the DSF balancing energy benefits based on the different marginal costs for eligible balancing
technologies at EU level. To do so, this approach was followed:

Step 1 consisted of data collection, DNV collected information on reserve volumes, reserve capacities and marginal
costs. The calculated total activated balancing reserves for EU 27 in 2021 (these are assumed to be the same in 2030).
The total activated reserves are extracted from ENTSO-E (ENTSO-E, 2022) and are summarised in the table below:

Reserve Volumes 2021 (GWh)

aFRR down -7735.64

aFRR up 8504.27

mFRR down -10746.07

mFRR up 8504.27

RR down -17689.66

RR up 13344.83

The reserve capacities were estimated based on a sample of requirements from 9 countries with relative lower
(Belgium) and higher (France) reserve requirements (ENTSO-E, 2022), (TERNA, 2020). The resulting range of required
balancing capacity for EU 27 is shown in the table below:

Reserve Volumes 2021 (GWh) Volumes 2021 (GWh)

LOW HIGH
RANGE RANGE

aFRR (upward/ downward) 12465 18450

mFRR (upward/ downward) 29760 45300

RR(upward/ downward) 6500 8500

44
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

The marginal costs for all technologies, depending on downward or upward reserve, are summarised in the tables
below:

Price €/ Calculation/Source
DOWNWARD aFRR mFRR RR MWH

Based on compensation to consumer that offers flex in

EV X X 2.9
ancillary services (Vandebron, 2017) (Jedlix, 2022)

Based on compensation to consumer that offers flex in

Heatpump X X 2.9
ancillary services (Vandebron, 2017) (Jedlix, 2022)

Average electricity price 2030 (output from market

EV – V2G X X X 9.7
model) multiplied by round-trip efficiency loss (13%)

Battery FTM X X X 9.7 Same as above

Battery BTM X X X 14.7 Same as above

Same as upward BatterLost revenue based on average

market price (output from market model) and saved
operational cost (gas price/efficiency + emission cost =
CCGT X X X 74.7
36.1 €/MWh). Gas, efficiency and emission prices are
based on the assumptions taken in the market model
for 2030M

Cost of lost revenue based on average electricity price

Wind curtailment X 74.7
in 2030 (output from the market model)

PV curtailment X 74.7 Same as above

Average electricity price in 2030/efficiency of pumping

Pumped hydro X X 99.6 (75%)

Electrolyzers X X 100.4 Average electricity price in 2030/efficiency (67.7%)

45
 Price €/ Calculation/Source
UPWARD aFRR mFRR RR MWH

EV X X X 3 Same as downward EV

Heatpump X X X 3 Same as upward heat pump

Pumped hydro X X X 6 Based on LCOE (C. Garcia Mazo, 2020)

EV – V2G X X X 10 Same as upward V2G

Battery FTM X X X 10 Same as upward Battery FTM

Battery BTM X X X 10 Same as upward Battery BTM

Based on gas price (25.3 €/MWth – market model),

CCGT X X X 60 gas energy density, efficiency (60%), emissions of gas
(201.9 kg/MWh LHV gas) and emission cost (53 €/ton)

CHP X X X 75 Same as above but considering 48% efficiency.

Same as cost of hydrogen in Appendix A – section

Electrolyzers X X 86 1.1.2.9

Gas turbines X X X 120 Same as CCGT but considering 30% efficiency.

Equivalent to Variable Operation & maintenance cost

Industrial DSR – 24 hr X X 207 of this flexibility. See Appendix 1 - section 1.1.2.1.

Equivalent to Variable Operation & maintenance cost

Industrial DSR – 4 hr X X 323 of this flexibility. See Appendix 1 - section 1.1.2.1.

Equivalent to bid price. See Appendix 1 – section

Industrial E-boiler X X 500 1.1.2.6.

Equivalent to Variable Operation & maintenance cost

Industrial DSR – 8 hr X X 500 of this flexibility. See Appendix 1 - section 1.1.2.1.

Equivalent to Variable Operation & maintenance cost

Industrial DSR – 1 hr X X 1000 of this flexibility. See Appendix 1 - section 1.1.2.1.

Equivalent to Variable Operation & maintenance cost

Industrial DSR – 2 hr X X 1097
of this flexibility. See Appendix 1 - section 1.1.2.1.

46
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Step 2 consisted of building the technology merit order for the provision of balancing energy in 2030 per scenario
(DSF and no-DSF scenario).

The DSF scenario included all technologies that are considered as DSF in the modelling exercise (e.g. smart charging,
batteries behind the meter, electric heating, industrial DSR, etc) as well as other flexible technologies (electrolysers
and front of the meter BESS) and traditional balancing reserve providers (gas turbines, gas power plants, hydro
power, etc.)

The capacities used to build the merit order were derived from the calculated DSF power in the previous section,
as well as the input generation capacities in DNV’s power market model for 2030.

An example of merit order is presented in Figure 12. The figure includes the technology merit order for providing
upward aFRR energy for each scenario. The green band represents the required aFRR capacity for all EU. Figure
13 zooms in the merit order to show more details in the relevant capacity range. It can be observed that all aFRR
upward capacity can be provided by hydro energy in the no-DSF scenario (at around 6 €/MWh) and by residential
DSF in the DSF scenario (at around 3 €/MWh).

Figure 12 – Upward aFRR merit order (all capacity)

Figure 13 – Upward aFRR merit order (zoom-in)

47
 Figure 14 – Merit order for all balancing energy services

48
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

Based on the merit order and the required capacities, the cut out marginal costs for balancing energy is determined.
For some products there is a range of marginal costs because the reserve required capacity band intersects with
more than one technology. The results are presented in the table below:

Marginal price 2030 DSF Marginal price 2030 no-DSF

UPWARD (€/MWh) (€/MWh)

Low High Low High

aFRR down 2.85 2.85 9.72 9.72

aFRR up 2.85 2.85 5.62 5.62

mFRR down 2.85 2.85 14.66 14.66

mFRR up 2.85 2.85 9.72 60.07

RR down 9.72 9.72 9.72 9.72

RR up 5.62 5.62 5.62 5.62

Finally, step 3 consisted of calculating the balancing costs for all reserve services based on the merit orders in step
2. The DSF benefits were calculated as the difference between the no-DSF scenario balancing costs and the DSF
scenario balancing costs. The results are summarised below:

UPWARD DSF balancing benefits (million €)

Low High
aFRR down 53.1 53.1

aFRR up 23.6 23.6

mFRR down 127.0 127.0

mFRR up 58.4 486.6

RR down 0.0 0.0

RR up 0.0 0.0

TOTAL 262.0 690.2

1.2.2. LIMITATIONS OF THE APPROACH

DNV acknowledges that the proposed approach for the quantification of system balancing benefits shows some
limitations which are briefly presented below:

The balancing requirements at European level are calculated as a sum of country-specific balancing-capacity
requirements. DNV does not quantify the synergies of having a unified European market and how that affects the
balancing requirements.

The quantification assumes that there is enough interconnection capacity available between bidding zones, so
all flexibility is available to all systems.

The merit order is built based on marginal costs of the different technologies only. Other costs such as opportunity
costs (e.g. the missed revenue for not selling their flexibility in other markets such as congestion management) are
not included.

49
 The required balancing volumes in 2030 are considered to be the same as 2021.

1.3. Quantification of grid infrastructure benefits

1.3.1. DESKTOP RESEARCH

DNV has conducted desktop research on the infrastructural benefits that DSF may enable. The research looked for
studies that have quantified the benefits related to delayed or deferred grid reinforcements by 2030, preferably
published in the last 5 years and focusing in one or multiple countries for the EU 27+ area. Table 1.1 provides a
selection of the studies reviewed by DNV for this desktop research.

Table 1.11 – Selection of studies reviewed by DNV for the desktop research on DSF infrastructural benefits

Year of
Reference Author Title publication

Cadrage des hypothèses sur les gisements

(RTE, 2020) RTE
de flexibilité de la demande
2020

(Energy Union Energy Union

Cleaner, Smarter, Cheaper 2017
Choices, 2017) Choices

(European European
Mainstreaming RES Flexibility Portfolios 2017
Commission, 2017) Commission

CO2 free flexibility options for the Dutch

(Aurora, 2021) Aurora
power system
2021

(ECN, Alliander, ECN, Alliander

The supply of flexibility for the power
2017
2017) system in the Netherlands, 2015-2050

Adequacy and flexibility study for Belgium

(Elia, 2021) Elia
2022-2032
2021

(RTE, 2021) RTE Energy pathways to 2050 2021

The balance of power –flexibility options for

(Triple, 2015) Triple
the Dutch electricity market
2015

(European Climate European Climate

Roadmap 2050 2010
Foundation, 2010) Foundation

(Arthur D. Little, Arthur D. Little

Distributed flexibility: the next pool of value,
2020
2020) 2020

Samhällsekonomiska kostnader och nyttor

(DNV GL, 2021) DNV GL
av smarta elnät
2021

Kostnader i strømnettet gevinster ved

(DNV GL, 2019) DNV GL
koordinert lading av elbiler
2019

(Carbon Trust, Carbon Trust Flexibility in Great Britain 2021

2021)

Connecting the dots: distribution grid

(Eurelectric, 2021) Eurelectric
investment to power the energy transition
2021

Wirtschaftlicher Vorteil der netzdienlichen

(E-Bridge, 2019) E-Bridge 2019
Nutzung von Flexibitaet in Verteilnetzen

50
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

The desktop research has shown that very little literature is currently available on the quantification of infrastructural
benefits in terms of delayed or deferred grid investments that DSF may enable. On the one hand, the interpretation
of the concept of flexibility in many studies is not consistent with the one used for this evaluation. Flexibility is
also often sought on the generation side of the system rather than on the demand side, for instance by means
of CCGT gas turbines. When focusing on DSF, many studies primarily quantify of the flexibility potential of one or
multiple technologies without investigating the potential infrastructural benefits. Others provide a very high-level
definition of DSF, for instance 20% of the peak demand that can be shifted throughout the day. On the other hand,
the results of those studies that quantify the benefits may not be applicable to the task at hand. For example, some
studies quantify the benefits of additional flexible technology capacity rather than the benefits of the use of flexible
technologies. Other results are too case-specific or grid-specific to be scaled at EU level.

Overall, DNV acknowledges the scarcity of studies on the infrastructural benefits of DSF. It was, therefore, impossible
to pursue the original plan of estimating the infrastructural benefits in a Fit for 55 scenario based on existing studies
based on either country-specific or EU 27 grid simulations.

DNV has, therefore, defined a two-phase approach to estimate the infrastructural benefits at EU level relying on the
limited number of relevant studies. Next, the approach is detailed, and its main limitations are discussed.

1.3.2. QUANTIFICATION APPROACH

DNV approach for the quantification of infrastructural benefits focuses on distribution grid investments and it is
structured in two phases: a top-down phase to estimate the required investments per country and a bottom-up
phase to estimate the potential savings at EU 27 level. The approach strongly relies on the figures presented in the
(Eurelectric, 2021) study on the required distribution grid investments in EU 27+ to enable the energy transition and
in the (E-Bridge, 2019) study on the benefits of the grid-friendly use of flexibility in German distribution networks.
The study (Eurelectric, 2021) assesses DSO investments requires for enabling the energy transition in Europe and
develops policies and recommendations. The study estimates 375 – 425 billion € of power distribution grid require
investments between 2020 and 2030 in EU 27+UK. These figures are extrapolated based on information provided
by the DSOs of 10 countries accounting for about 70% of EU electricity demand. The required investments are
divided into four main areas, the largest being due to electrification and renewables, and eight main investment
drivers. DNV refers to this study to estimate the required investments in distribution grids by 2030 in a Fit for 55
scenario based on three main drivers, namely final electricity demand, electric vehicles, and RES capacity connected
to distribution grid.

The study (E-Bridge, 2019) quantifies the economic potential of flexibility for congestion management in the German
distribution grid for the year 2035 and determines the possible benefits in terms of deferral of grid expansion
investments through a better utilisation of the grid. The study relies on thousands of grid simulations and shows
that, while the increase in flexibility used purely for the benefit of the market/system leads to an increasing need
for grid expansion, a grid-friendly use of flexibility may reduce the required investments in German distribution
grid by 55%. DNV refers to this study to estimate the potential savings in grid investments at EU 27 level thanks to
a grid-friendly use of demand side flexibility.

51
 The approach proposed by DNV is shown in Figure 15.

1 Eurelectric 2021 6
EU27+UK EU27 savings in
investments per fit-for-55 scenario
driver

2 5
EU27 investments % savings per
based on fit-for-55 country in
scenario fit-for-55 scenario

3 4
Required EBridge2019 DSF
investments per power 76.9%
country savings

Figure 15 – DNV’s approach for quantification of infrastructural benefits

1.3.2.1. Top-down phase

The first phase of the DNV approach is top-down and The top-down phase consists in the following steps:
estimates the required investments in distribution
grid between 2023 and 2030. It uses the figures from 1 Estimate of the investment per unit of driver based
(Eurelectric, 2021) on the investments in distribution grids on (Eurelectric, 2021). The study (Eurelectric, 2021)
due to electrification and renewables and its breakdown estimates that the electrification of the energy demand
per relevant investment drivers, namely, electrification and the increase in renewable capacity will require 180–
of buildings and industry, electrification of mobility, and 210 billion € investments in distribution grids between
emission-free generation. 2020 and 2030. The main drivers are the following:
increase in final electricity demand, electric vehicles, and
The bottom-up phase estimates the infrastructural benefits emission-free generation. The study expects an increase
in 2030 in a Fit for 55 scenario thanks to DSF. It uses the in total power demand between 2017 and 2030 of 732
figures from (E-Bridge, 2019) on the potential savings in TWh, which is responsible for 70–80 billion € investment.
investments when DSF is operated in a grid-friendly way Additionally, the number of EVs in 2030 is expected to
rather than in a market/system friendly way. Based on reach 50–70 million vehicles and is responsible for 25–35
the market simulation results, the savings presented by billion €. Furthermore, the renewable capacity connected
(E-Bridge, 2019) for German distribution networks are to distribution grids is estimated to increase by 360 GW
scaled and applied to the other countries. Finally, the between 2017 and 2030, requiring 85–95 billion €. The
total EU 27 savings are calculated. impact of each driver on the total required investments
is summarised in Table 1.12.

Table 1.12 – Impact of investment drivers on the required investments in distribution grid due to electrification
of the energy demand and increase in renewable capacity according to (Eurelectric, 2021)

Driver Quantifying parameter Investment per unit of driver

Electrification of demand Increase in total demand 0.096–0.109 bn/TWh

Electrification of mobility Number of EVs 0.5 bn/million

RES capacity addition to

Emission-free generation 0.236–0.264 bn/GW
distribution grids

52
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

2 Estimate of required investments in distribution grids in EU 27 between 2023 and 2030 in Fit for 55 scenario. Using
the DNV’s European Power Market Model for 2023 and the 2030 model developed for this study, DNV estimates the
values of investments associated with the main investment driver identified at the previous step. The share of solar
PV and wind capacity connected to the distribution grid is calculated using the figures provided for ten countries
by (Eurelectric, 2021). For those countries for which the study does not report a figure, the share of countries with
comparable area is used. Table 1.13 summarizes the results. The total required investments in distribution grids in
EU 27 between 2023 and 2030 are between 253.1 and 282.5 billion €.

Table 1.13 – Required investments in distribution grid between 2023 and 2030 in Fit for 55 scenario

Driver Quantifying parameter Quantity EU 27 Fit for 55 investment

Electrification Increase in total demand 1352 TWh 129.3–147.8 billion

of demand

Electrification Number of EVs 61 million 30.5 billion

of mobility

Emission-free RES capacity addition to

395 GW 93.3 – 104.3 billion
generation distribution grids

3 Estimate required investments in distribution grids per country between 2023 and 2030. The estimated investments
at EU 27 level are distributed across countries using three metrics that quantify the extent to which each country
contributes to the investment drivers at European level previously identified. Namely, the metrics represent the share
of increase in total demand, the share in EVs, and the share of additional RES capacity. The range of investments
253.1 – 282.5 billion € is, hence, distributed cross countries using these metrics. As a result, six values are obtained
per country and a range of investment is defined. See box below for an illustrative example.

EU 27 investments:
253.1 – 282.5 billion €

Country X:
Share in increase in total demand: 3% (40.6 TWh)
Share in EVs: 1.8% (1.1 million)
Share in additional RES capacity: 4% (15.9 GW)

Country X range of investments: 4.7 – 11.4 billion

53
 1.3.2.2. Bottom-up phase
The bottom-up phase consists of the following steps:

4 Calculate relation between available flexible capacity and savings. The study (E-Bridge, 2019) estimates the
flexible capacity coming from electric heat applications at 6.5 GW, from EVs at 10.5 GW, and from small storage
at 5.5 GW. Additionally, it estimates that residential, industrial, and trade and commerce traditional load can
provide 12%, 4.8%, and 21% of flexibility, respectively. Since the study reports no figure of the peak load and it was
conducted in 2019, the German 2035 estimate for peak load according to the 2019 update of the DNV’s European
Power Market Model is used as a reference, i.e., 133 GW of peak demand in 2035. Averaging the flexibility ratios
of traditional load reported by the study, it can be assumed that 12.6% of the peak demand is flexible. In total, the
calculated available flexible capacity by 2035 is 39.3 GW, which results from adding up flexibility of EVs, storage,
e-heating and flexible traditional load.

Therefore, it is calculated that a flexible capacity to peak demand of 29.5% leads to 76.9% savings.18

5 Estimate savings in % per country. The potential percentage savings are calculated per country based on:

Available flexible capacity;19

the relation derived at step 4, i.e., a flexibility capacity to peak load ratio of 29.5% leads to 76.9% savings;
and
peak demand in 2030.

The savings in % are calculated per country as follow:

availableFlexCapacity 76.9%
savings%=
peak load 29.5%

See box below for an illustrative example.

Country X – market simulation output:

Available flexible capacity: 2 GW

Peak load: 9.9 GW

Step 5:
Country X savings in %: 52.6% (2/9.9 * 76.9%/29.5%)

18. This percentage of savings only accounts for savings on low voltage and medium voltage distribution grids
54 19. DSF flexible power is derived from the results of section 2 – quantification of DSF in 2030. The grid benefits quantification approach is based on the upward flexible power.
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

6 Estimate savings at EU 27 level. Based on the estimated savings in % per country according to step 5, the
potential range of savings in each country is calculated. Finally, the potential infrastructural savings in distribution
grid investments at EU 27 level between 2023 and 2030 thanks to grid-friendly DSF are calculated by summing up
the country-specific values.

The savings at EU 27 level between 2023 and 2030 are estimated at 77.6–203.6 billion €, which corresponds to
annual savings of 11.1 and 29.1 billion €.

1.3.3. LIMITATIONS OF THE APPROACH

DNV acknowledges that the proposed approach for the quantification of infrastructural benefits shows some
limitations, which are briefly presented below:

The relation between DSF power and investment savings The approach does not take into account the differences
is neither derived from literature nor from grid model in the current status of development across distribution
simulation results, but rather calculated based on data grids. For example, those countries still mostly relying
derived from two different sources, namely, (E-Bridge, on gas-based heating systems might face larger costs to
2019) and DNV’s European Power Market Model 2019 enable distribution grids to cope with increasing heat
update. pump capacity and electric heating load than those
countries already largely relying on electric heating
The quantified investments and savings refer to the systems.
distribution grid only. Investments required in transmission
grids are not quantified. The approach does not take into account the differences
across countries regarding voltage level in the distribution
The quantified investments account for integration grid. The approach assumes that the savings are on low
of new (electrified) loads and RES capacity only. Other voltage and medium voltage distribution grid.
investments, such as digitalisation and smart meter roll
out, are not calculated. The investment costs per driver are assumed to be
linear while they may strongly depend on other factors,
such us the current status of the distribution grids.

55
 APPENDIX B – References

Arthur D. Little. (2020). Distributed flexibility: the next pool of value, 2020.
Aurora. (2021). CO2 free flexibility options for the Dutch power system.
BMWK. (2022). Germany, Denmark, Netherlands and Belgium sign €135 billion offshore wind pact. Retrieved
from: https://www.bmwk.de/Redaktion/DE/Downloads/Energie/20220518-declaration-of-energy-ministers.
pdf?__blob=publicationFile&v=10
California Institute of Technology. (2021). ACN-Data. Retrieved from https://ev.caltech.edu/dataset
Carbon Trust. (2021). Flexibility in Great Britain.
DNV. (2021). Energy Transition Outlook (ETO) 2021.
DNV GL. (2019). Kostnader i strømnettet gevinster ved koordinert lading av elbiler.
DNV GL. (2021). Samhällsekonomiska kostnader och nyttor av smarta elnät .
EASE, D.-E. (n.d.). EMMES 6.0. Retrieved 2022, from https://ease-storage.eu/publication/emmes-6-0-june-2022/
E-Bridge. (2019). Wirtschaftlicher Vorteil der netzdienlichen Nutzung von Flexibitaet in Verteilnetzen.
ECN, Alliander. (2017). The supply of flexibility for the power system in the Netherlands, 2015-2050.
ElaadNL. (2020). Open EV charging datasets. Retrieved from https://platform.elaad.io/analyses/index.php?url=ElaadNL_
opendata.php
Elia. (2019). Adequacy and flexibility study for Belgium 2020-2030.
Elia. (2021). Adequacy and flexibility study for Belgium 2022-2032.
Energy Union Choices. (2017). Cleaner, Smarter, Cheaper.
ENTSO-E. (2020). Ten-Year Development Plan 2020.
ENTSO-E. (2021). European Resource Adequacy Assessment (ERAA).
ENTSO-E. (2022). Balancing report.
ENTSOG and ENTSO-E. (2022). Ten-Year Network Development Plan (TYNDP) 2022 - Scenario Report - Version April
2022.
Eurelectric. (2021). Connecting the dots: distribution grid investment to power the energy transition.
European Climate Foundation. (2010). Roadmap 2050.
European Commission . (2021). ‘Fit for 55’: delivering the EU’s 2030 Climate Target on the way to climate neutrality.
European Commission. (2017). Mainstreaming RES flexibility portfolios.
European Commission. (2022). REPowerEU Plan.
European Commission DG Energy. (2016). Impact assessment study on downstream flexibility, price flexibility, demand
response & smart metering. European Commission DG Energy.

56
 DEMAND-SIDE FLEXIBILITY IN THE EU: Quantification of benefits in 2030

European Commission, Directorate-General for Energy. (2022). Best practices and assessment of regulatory measures
for cost-efficient integration of electric vehicles into the electricity grid. Publications Office of the European Union.
eurostat. (2020, 07 10). Retrieved from https://ec.europa.eu/eurostat/documents/2995521/11081093/3-10072020-
AP-EN.pdf/d2f799bf-4412-05cc-a357-7b49b93615f1
Jedlix. (2022). Retrieved from https://www.jedlix.com/nl/mountox-landingspage/
LAZARD. (2020). Lazard’s Levelized Cost of Energy Analysis - version 14.0.
Planbureau voor de Leefomgeving. (2020). EINDADVIES BASISBEDRAGEN SDE++ 2020.
RTE. (2016). ÉVALUATION DES ÉCONOMIES D’ÉNERGIE ET DES EFFETS DE BORD associés aux effacements de
consummation.
RTE. (2020). Cadrage des hypothèses sur les gisements de flexibilité de la demande.
RTE. (2021). Energy pathways to 2050.
Ruhnau, O. H. (2019). Time series of heat demand and heat pump efficiency for energy system modeling.
TERNA. (2020). Terna report on Balancing in accordance with Article 60 of COMMISSION REGULATION (EU) 2017/2195
of 23 November 2017 establishing a guideline on electricity balancing.
Thomas D. (Hydrogenics), M. D. (2016). POWER TO GAS Roadmap for Flanders.
Triple. (2015). The balance of power –flexibility options for the Dutch electricity market.
UK Government. (2022). Electric vehicle charging transaction. Retrieved from https://www.data.gov.uk/dataset/16c7326b-
57fe-4803-88f8-9286c387f68a/electric-vehicle-charging-transactions
Vandebron. (2017). Retrieved from Slim laden: hoe werkt dat?: https://vandebron.nl/blog/slim-laden-hoe-werkt-dat

57
COPYRIGHT
Copyright © 2022 smartEn Smart Energy Europe. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form
or by any means electronic, mechanical, photocopying, recording or otherwise without the prior
written permission of Smart Energy Europe.

Unless otherwise credited all diagrams in this report belong to Smart Energy Europe.

IMPORTANT
This document contains confidential and commercially sensitive information. Should any requests for
disclosure of information contained in this document be received, we request that we be notified in
writing of the details of such request and that we be consulted and our comments taken into account
before any action is taken.

DISCLAIMER
While Smart Energy Europe and DNV consider that the information and opinions given in this work
are sound, all parties must rely upon their own skill and judgement when making use of it. Smart
Energy Europe and DNV do not make any representation or warranty, expressed or implied, as to the
accuracy or completeness of the information contained in the report and assumes no responsibility
for the accuracy or completeness of such information. Smart Energy Europe and DNV will not assume
any liability to anyone for any loss or damage arising out of the provision of this report.

Where this report contains quantifications, these are based on assumptions that are subject to
uncertainties and contingencies. Because of the subjective judgements and inherent uncertainties of
projections, and because events frequently do not occur as expected, there can be no assurance that
the projections contained herein will be realised and actual events may be difference from projected
results. Hence the projections supplied are not to be regarded as firm predictions of the future, but
rather as illustrations of a modelling exercise.
Platinum Partners

Eaton is an intelligent power management company dedicated to improving

the quality of life and protecting the environment for people everywhere. We
are guided by our commitment to do business right, to operate sustainably
and to help our customers manage power ─ today and well into the future. By
capitalizing on the global growth trends of electrification and digitalization, we’re
accelerating the planet’s transition to renewable energy, helping to solve the
world’s most urgent power management challenges, and doing what’s best for
our stakeholders and all of society.

Enel is a multinational energy company and one of the world’s leading integrated
utility within the energy sector. We are present in more than 40 countries across
five continents, generating energy with a total installed capacity of around 92
GW (of which around 55 GW of renewables capacity), selling electricity and gas
and distributing electricity across a network spanning more than 2.2 million
km. With almost 76 million end users worldwide, we are the biggest networks
operator, we manage the largest customer base in the utilities sector with around
70 million retail customers and we are one of Europe’s leading energy companies
by installed capacity.

The Enel Group counts on a workforce of more than 67,000 people from around
the world and its foundations are based on the values of Responsibility, Innovation,
Trust and Proactivity.

Gold Partner
As a major player in energy transition, the EDF Group is an integrated energy
company active in all businesses: generation, transmission, distribution, energy
trading, energy sales and energy services. EDF group is a world leader in low-
carbon energy, having developed a diverse production mix based mainly on
nuclear and renewable energy (including hydropower). It is also investing in
new technologies to support energy transition. EDF’s raison d’être is to build
a net zero energy future with electricity and innovative solutions and services,
to help save the planet and drive well-being and economic development. The
Group is involved in supplying energy and services to approximately 38.5 million
customers , of whom 29.3 million in France. It generated consolidated sales of
€84.5 billion in 2021. EDF is listed on the Paris Stock Exchange.

Silver Partner