DISTRIBUTION SYSTEM OPERATORS

OBSERVATORY

From European Electricity Distribution Systems to

Representative Distribution Networks

Giuseppe PRETTICO, Flavia GANGALE,

Anna MENGOLINI, Alexandre LUCAS,
Gianluca FULLI

2016

EUR 27927 EN
 DISTRIBUTION SYSTEM
OPERATORS OBSERVATORY

From European Electricity Distribution

Systems to Representative Distribution
Networks
This publication is a Technical report by the Joint Research Centre, the European Commission’s in-house science
service. It aims to provide evidence-based scientific support to the European policy-making process. The scientific
output expressed does not imply a policy position of the European Commission. Neither the European
Commission nor any person acting on behalf of the Commission is responsible for the use which might be made
of this publication.

Contact information
Name: Giuseppe Prettico
Address: Via E. Fermi 2749 TP 693 I-21027 Ispra (VA) Italy
E-mail: [email protected]
Tel.: +39-0332-783675

JRC Science Hub

https://ec.europa.eu/jrc

JRC101680

EUR 27927 EN

ISBN 978-92-79-58416-9 (PDF)

ISBN 978-92-79-58415-2 (print)

ISSN 1831-9424 (online)

ISSN 1018-5593 (print)

doi: 10.2790/701791 (online)

doi: 10.2790/471701 (print)

© European Union, 2016

Reproduction is authorised provided the source is acknowledged.

Printed in Luxembourg

All images © European Union 2016, except: Cover page: Photocase, Korre – Sparkling City II

How to cite: Prettico G., Gangale F., Mengolini A., Lucas A. and Fulli G. ; DISTRIBUTION SYSTEM OPERATORS
OBSERVATORY: From European Electricity Distribution Systems to Representative Distribution Networks; EUR
27927 EN; 10.2790/701791
TABLE OF CONTENTS
Acknowledgements ................................................................................................ 1
Executive Summary............................................................................................... 3
1 Introduction .................................................................................................... 7
2 State of play of Distribution System Operators in Europe ...................................... 9
2.1 DSOs’ Observatory Project ........................................................................ 11
3 A clearer view of the European distribution networks ......................................... 15
3.1 Survey participation ................................................................................. 15
3.2 DSOs indicators ....................................................................................... 18
3.2.1 Main DSOs indicators explained ............................................................ 21
4 From Distribution Systems to Representative Distribution Networks ..................... 29
4.1 The Reference Network Model concept ........................................................ 29
4.1.1 RNM Outputs ..................................................................................... 30
4.1.2 RNM Inputs........................................................................................ 30
4.1.3 RNM planning steps ............................................................................ 32
4.1.4 RNM Methodology ............................................................................... 32
4.2 Realistic synthetic networks based on RNM .................................................. 37
4.2.1 Representative large-scale networks ..................................................... 39
4.2.2 Feeder-type network topologies ........................................................... 48
5 Evaluation of policy options based on RNM ........................................................ 65
5.1 Selected scenarios ................................................................................... 66
5.2 Results of the simulations ......................................................................... 67
5.2.1 Impact of solar PV on distribution networks ........................................... 68
5.2.2 Storage units to mitigate voltage spreads in the network ......................... 75
5.2.3 Reliability analysis .............................................................................. 77
6 Conclusions .................................................................................................. 81
Annex A: Indicator box plots ................................................................................ 83
Annex B: Other indicators .................................................................................... 89
Annex C: On-line survey .................................................................................... 101
Annex D: Cost functions ..................................................................................... 105
References ....................................................................................................... 107
List of abbreviations and definitions ..................................................................... 111
List of figures.................................................................................................... 115
List of tables ..................................................................................................... 117
Acknowledgements
We are very grateful to the IIT from Comillas Pontifical University for the fruitful
collaboration established within the Distribution System Operators Observatory Project.
Special thanks go to Carlos Mateo, Tomás Gómez, Rafael Cossent and Pablo Frías. We
also acknowledge the key support offered by EURELECTRIC, in particular by Jacobo
Alvarez and Susan Nies, during the data collection process.

1
2
Executive Summary

The European electricity sector is undergoing radical changes in every segment of the
power industry, from generation to supply. Ambitious policy goals set at European level
to enhance the competitiveness, security and sustainability of the EU's energy system
have called for major changes in the regulatory, technological, and market structure
fields.

The distribution sector is particularly affected by these changes. The increasing

penetration of local renewable generation and the emergence of demand response
enabling solutions are placing new requirements on the distribution networks, posing
challenges to the reliability and efficiency of system operation. At the same time,
however, these new applications can also create opportunities to manage the distribution
grids in a more flexible and efficient way, paving the way to new services to end-
consumers.

Smart grid concepts and technologies have an important role to play to address these
new challenges and opportunities. A variety of solutions are already being tested in
Europe, with encouraging results. Identifying the best technical, economic and societal
options requires a deep understanding of their impact on the physical distribution
networks. Such knowledge is also necessary to evaluate the viability of replicating and
scaling up pilot experiences already successfully implemented in Europe.

At present, there is little publicly available information on the European distribution

system operators (DSOs) and the networks they operate. This lack of knowledge is
partially attributable to constraints in sharing data that DSOs consider as assets of
commercial value, but it is also due to the vast number and heterogeneity of the
distribution systems in Europe. The situation varies radically from country to country,
due to historical as well as geographical, legal, political and economic reasons. In some
Member States there is only one DSO, while in others there are tens or hundreds of
them operating their networks on a regional or even municipal basis. Differences
concern also other aspects, e.g. the scope of the DSO activities, the level of unbundling,
the operated voltage levels and other key technical information on the networks.

In the last years, the Joint Research Centre of the European Commission (JRC) has
expanded its role as an independent observer of the energy system. A big effort has
been put in collecting, processing and analysing data on the power sector (from smart
grid project costs and benefits to consumer engagement strategies, from power system
techno-economic features to integrated regional systems/markets). This activity is aimed
at providing stakeholders with tools and analyses to better understand the rapidly
changing scene, enabling early identification of developments and opportunities and
supporting evidence-based policy making.
This report presents the latest JRC data brokering effort – the DSO Observatory project -
focused on European distribution system operators and their distribution networks. The
aim of the report is to contribute to a better understanding of the challenges that the
transition to a new energy system is posing to European distribution system operators
and to elaborate sound solutions to address them.

3
To the best of our knowledge, this is the most comprehensive data collection exercise on
European distribution systems published so far. Based upon this inventory, detailed
Reference Network Models (RNMs) were used to develop representative distribution
networks for analysing the impact that Renewable Energy Sources (RES) penetration and
network automation can have on their technical performance.

The starting point of the DSO Observatory project was the collection of technical and
structural data from the DSOs. Given the vast number of DSOs in Europe, the data
collection exercise was limited to the bigger ones, i.e. the 190 DSOs that have to comply
with the unbundling requirements set out in the EU Electricity Directive (i.e. the DSOs
serving more than 100,000 customers, also referred to in the report as "larger" DSOs).
An online survey was launched in January 2015 with the aim of collecting several
clusters of data, relating to types of ownership and unbundling, network structures and
designs, amounts and types of connected distributed generation, and reliability of supply
indicators.

79 out of the 190 larger DSOs responded to the survey. The representativeness of the
obtained sample is quite high: the respondents manage more than 70% of the electricity
supplied by all DSOs serving over 100,000 customers. Overall, the 79 DSOs distribute
more than 2,000 TWh of electricity to over 200 million customers per year, covering a
total area of more than 3 million square km.

The collected data were used to build 36 indicators, divided in three categories, i.e.
network structure, network design and distributed generation. These indicators allow for
comparison of the parameters and criteria used by DSOs when designing and sizing their
network installations. They help to shed some light on the different characteristics of
some of the major European distribution networks and to support research activities by
reducing the amount of resources that are typically devoted to compiling input data and
building case studies.

The project also aimed at providing a tool to enable more sophisticated technical and
economic assessments of different policy and technological solutions. For this purpose,
10 of the 36 indicators were chosen to create representative distribution networks using
RNMs, i.e. large-scale network distribution planning tools that allow designing realistic
distribution networks useful for simulation activities. RNMs allow the design of networks
that supply the expected demand while taking into consideration the need to minimize
the total investment and associated operational costs and to meet the defined reliability
of supply criteria. By providing a realistic distribution network, RNMs offer the possibility
to reliably simulate the impact of different scenarios on the grid without the need to
have access to the actual network data.

Two large-scale representative networks, one rural and one urban, were selected to
carry out the simulation analyses. The networks were used to analyse the impact of
increasing levels of RES penetration, in particular solar photovoltaic (PV) and wind, on
the technical performance of the grid. The impact on network voltages and network
overloads was then monetised by means of a penalty cost function.

The analysis shows that the number and size of PV units, as well as their
concentration/distribution on the network, are all relevant parameters. By way of
illustration, results highlight that limiting the size of the generation units would allow
mitigating voltage and congestion problems, maximizing renewable penetration with no

4
need of additional network investments. A careful consideration of the local conditions of
each distribution area as well as of the different connection patterns - including unit
sizes, technologies and location within the network - is therefore of paramount
importance to minimize adverse impacts on the network.

Another way to mitigate the voltage spread introduced by the increasing penetration of
PV connected to the distribution network has been studied considering the installation of
storage units where the PV units are located. The analysis highlights that significant
voltage spread reduction is only observed when big storage units are installed in
combination with each PV unit. The high costs per kWh of batteries (estimated on the
basis of the current storage solutions market) suggest that other solutions should be
considered to mitigate voltage spread in the distribution network.

Finally, a reliability analysis is presented, showing how the System Average Interruption
Frequency Index (SAIFI) and the System Average Interruption Duration Index (SAIDI)
could be reduced by increasing the installation of tele-controlled switches in the
distribution network.

The analyses presented in the report provide an illustrative example of the potential
applications of the representative networks built within the DSO Observatory project.
Other applications are however possible and this report can be seen as a first step of an
exercise that will continue in the future. To foster more research on such a key subject,
the JRC is planning to share the representative networks built as part of the DSO
Observatory project with all interested parties carrying out research activities and
techno-economic studies in this field1. The JRC will continue and expand its scientific and
policy support activities in this sector to better understand and address the challenges
DSOs face in the transition to the new energy system.

1
To get the built representative networks modeled in Matlab/Matpower please visit the webpage:
http://ses.jrc.ec.europa.eu/distribution-system-operators-observatory

5
1 Introduction

In the last years the European electricity sector has gone through radical changes in
every segment of the power industry, from generation to supply. Ambitious policy goals
set at European level to enhance the competitiveness, security and sustainability of the
EU's energy system have called for major changes in the regulatory, technological, and
market structure fields.
All actors of the power industry are affected by these changes. Until recently, much
research and debate has occurred over the bulk power system (generation and
transmission), while less attention has been paid to the planning, operation, and
management of the distribution systems (L'Abbate, Fulli, Starr, & Peteves, 2008).

Lately however, the rapid changes occurring in the distribution segment have brought
this sector at centre stage of the debate. The increasing penetration of local renewable
generation and the emergence of demand response enabling solutions are acting as
main transformative forces in the power sector, making the distribution grids the
primary recipient of all the new interactions initiated by these numerous distributed units
(Glachant, Rious, & Vasconcelos, 2015). These new technologies are expected to
radically change local electricity industry and markets at the distribution level (Ruester,
Perez-Arriaga, Schwenen, Batlle, & Glachant, 2013), creating opportunities but also
posing challenges to the reliability and efficiency of system operation.

Several smart grid solutions are being tested in Europe to address these challenges and
opportunities. Identifying the best solution requires a deep understanding of the impact
of the different options on the distribution networks. Technical and economic
assessments of alternative investments require a detailed knowledge of the networks
and of the context in which they are operated. Such knowledge is also necessary to
evaluate the viability of replicating and scaling up pilot experiences already successfully
implemented in Europe.

Yet, not much is known about European distribution system operators (DSOs) and the
networks they operate. This lack of knowledge is due to the confidential nature of much
of this information as well as to the vast number and heterogeneity of electricity
distribution systems in Europe. According to a recent study (Eurelectric, 2013), there are
well over 2,000 distribution system operators in Europe, connecting 260 million
customers, operating 10 million km of lines, and supplying 2,700 TWh of energy per
year.

The situation varies radically from country to country, due to historical as well as
geographical, legal, political and economic reasons. In some Member States there is only
one DSO, while in others there are tens or hundreds of them operating their networks on
a regional or even municipal basis. Differences concern also other aspects, e.g. the
scope of the DSO activities, the level of unbundling, the operated voltage levels and
other key technical information on the networks.

Very little information is available in the public domain and it generally does not allow for
in-depth analyses. Against this background, the Joint Research Centre decided to launch
an initiative – the DSO Observatory project - to fill in this knowledge gap by collecting a
variety of data directly from the distribution system operators. Given the huge number

7
of DSOs in Europe however, the collection exercise was restricted to the main ones, i.e.
those that have to comply with the unbundling requirements set in the EU Electricity
Directive (i.e. the DSOs serving more than 100.000 customers).
The data gathering exercise was the starting point for the creation of representative
distribution networks using Reference Network Models (RNM) that can reliably simulate
the impact of different scenarios on European distribution networks without the need to
have access to the actual technical data. For this purpose, on the basis of the collected
information, several network structural indicators were derived and later used to validate
the soundness of the representative distribution networks. These networks were then
converted into Matlab/Matpower format and used to analyse the impact of increasing
levels of RES penetration and network automation on the technical performance of the
grid.

This ambitious project aims at shedding some light on the different characteristics of
some of the major European distribution networks and at supporting research activities
by reducing the amount of resources that are typically devoted to compiling input data
and building case studies. This in turn will contribute to a better understanding of the
challenges that the transition to a new energy system is posing to European distribution
system operators and to elaborate sound solutions to address them.

8
2 State of play of Distribution System Operators in
Europe

Electricity distribution is a crucial activity in the value chain from production to

consumption of electricity, as it links the transmission grid with final customers.
Traditionally, the European electricity sector was dominated by vertically integrated
monopolies that were either state-owned or privately-owned. The primary components
of electricity supply - generation,
transmission, distribution, and supply DIRECTIVE 2009/72/EC
– were therefore integrated within
Distribution means the transport of electricity on
individual electric utilities.
high-voltage, medium-voltage and low-voltage
During the 1990s, the European distribution systems with a view to its delivery to
customers, but does not include supply.
Union and the Member States
decided to open the electricity Distribution system operator (DSO) means a
natural or legal person responsible for operating,
markets to competition in a gradual
ensuring the maintenance of and, if necessary,
way. The liberalization process developing the distribution system in a given area
intended to achieve competitive and, where applicable, its interconnections with
prices through the game of market other systems and for ensuring the long-term
ability of the system to meet reasonable demands
forces and to establish a unified
for the distribution of electricity.
energy market that would also be
conducive to ensuring secure energy
supplies (Glachant, Rious and
Vasconcelos 2015).

One of the main cornerstones of the reform was the separation between the competitive
parts of the industry (generation and supply) and the non-competitive parts
(transmission and distribution networks). Electricity networks were considered non-
competitive as, from an economic perspective, they are 'natural monopolies', implying
that alternative and competing networks can only be built at very high costs.

Successive rounds of EU electricity market legislation, lastly Directive 2009/72/EC,

introduced unbundling requirements, which oblige Member States to ensure the
separation of vertically integrated energy companies. These reforms resulted in the
separation of the various stages of energy supply, i.e. generation, distribution,
transmission and supply. While transmission and distribution were exempted from
competition and subject to regulatory control, generation and supply, as potentially
competitive activities, were open to liberalization.
As stated in the preamble of Directive 2009/72/EC, the reasoning behind the new rules
was that 'without effective separation of networks from activities of generation and
supply (effective unbundling), there is an inherent risk of discrimination not only in the
operation of the network but also in the incentives for vertically integrated undertakings
to invest adequately in their networks'.

9
For the distribution segment, the unbundling
requirements do not create an obligation to Total no. DSOs >
separate the ownership of assets of the Country of DSOs 100,000
(in 2011) customers
distribution system operator from the
vertically integrated undertaking, but provide Austria 138 13
for separation at functional and legal level Belgium 24 15
(article 26 of Directive 2009/72/EC). Bulgaria 4 3
Croatia n/a 1*
Functional unbundling requires that the
Cyprus 1 1
network operator is independent in terms of
Czech Republic 3 3
its organization and decision making rights Denmark 72 6
from the other activities not related to that Estonia n/a 1
network. Legal unbundling requires that the Finland n/a 7
distribution activities are done by a separate France 158 5
'network' company (separate legal entity); Germany 880 75
the network company must not necessarily Greece n/a 1
own the network assets but must have Hungary 6 6
'effective decision making rights' in line with Ireland 1 1
the requirements of functional unbundling Italy 144 2
(DG Energy & Transport 16.1.2004). Latvia 11 1
Lithuania 1 1
In other words, functional unbundling is a
Luxembourg 6 1
prerequisite in order to ensure the
Malta n/a 1*
independence of network operators in terms
Netherlands 11 8
of organization and decision making
Poland 184 5
processes, while legal unbundling involves the
Portugal 13 3
setting up of a separate network company
Romania n.a. 8
(Ropenus, et al. 2009).
Slovakia 3 3
Member States were left free to exempt from Slovenia n/a 1
these requirements those integrated Spain n/a 5
electricity undertakings serving less than Sweden 173 6
100,000 connected customers, or serving United Kingdom 7 7
small isolated systems. This last provision is
quite relevant, as the majority of electricity Table 2-1 DSOs number per Country
(Eurelectric 2013)
distribution system operators in Europe are * Data for Malta and Croatia were not
below the threshold set by the Electricity reported in Eurelectric (2013)
Directive (Table 2-1) and can therefore be
exempted from the EU unbundling
requirements.

According to (European Commission 2012), the majority of Member States have actually
made use of the exemption rule. In its last review on the status of the unbundling
requirements for DSOs, (CEER 2015) estimated that only 189 of the 2,400 DSOs
operating in Europe have been unbundled 2 . This picture is however quite fluid and
constantly changing, due to the ever changing national circumstances.

2
Different sources (e.g. (Eurelectric 2013), (European Commission 2012), (CEER 2013)) report different
figures for the number of DSOs in Europe. Such differences are mainly due to the different timing of the

10
Finally, the extent of the unbundling also changes from country to country, with the
more extensive form of separation, i.e. ownership unbundling, being adopted only in a
minority of cases.
As for the role of DSOs, it also varies from country to country, due to their heterogeneity
and to differences in national regulation. The traditional role of DSOs is to operate,
maintain and develop the distribution network to ensure that electricity is delivered to
end-users in a secure, reliable and efficient manner. DSOs also play a role in the
efficient functioning of the European electricity markets, as they act as “entry gates” to
retail markets in most EU countries (CEER 2013), potentially influencing the level of
competition in this segment. They should therefore guarantee non-discriminatory access
to the grid and provide system users with the information they need for efficient access
to, including use of, the system.

Lately however, the changes triggered by the increasing penetration of local renewable
generation and by the emergence of demand response enabling solutions are calling for
a reconsideration of the role of DSOs. DSOs will be increasingly required to perform
more (pro-) active grid development, management and operation as these changes
place new requirements on the networks in terms of operational security, while they
offer at the same time more options for the DSOs to manage their grids in a more
flexible and efficient way (van den Oosterkamp, et al. 2014).

Research and debate is still open with respect to the new tasks, responsibilities and
opportunities that are shaping up in the evolving power system (CEER 2015) (Eurelectric
2010) (Ruester, et al. 2013) (van den Oosterkamp, et al. 2014). These new tasks could
in principle be performed by DSOs or they could be open to new and competing actors,
in a market environment.

As highlighted above, European DSOs differ in size and activity profile, as well as in the
technical characteristics of the networks and the challenges they need to face. Finding a
common European approach to DSO regulation is therefore a challenging task (CEER
2015).

2.1 DSOs’ Observatory Project

In the last years, the JRC has played an increasing role in collecting, processing and
analyzing data on the power sector. This activity is aimed at providing different
stakeholders with instruments to better understand the rapid changing scene, enabling
early identification of developments and opportunities and supporting evidence-based
policy making.

The first field which was investigated was that of smart grid research and development
(R&D) and demonstration activities. In 2011, the first Smart Grid Projects Outlook,
containing data and insights on smart grid projects in Europe, was released (Giordano,

surveys and the fluidity of the sector. In the rest of this report, unless otherwise stated, we will refer to
(Eurelectric 2013).

11
Gangale, et al. 2011). The report was updated twice, in 2013 (Giordano, Meletiou, et al.
2013) and in 2014 (Covrig, et al. 2014). In 2015, the data brokering activity was
extended to the European laboratories landscape (Poncela, et al. 2015). The idea was to
get an overview of all the smart grid technologies operational at laboratory level and to
identify research activities, gaps and future trends.
In this context, the DSOs Observatory project can be seen as a further extension of the
JRC data collection and analysis efforts to cover one of the main actors of the evolution
towards a new electricity system, i.e. the distribution system operators.

The DSO Observatory project was launched at the end of 2014. The project had three
main objectives: 1) getting an overall picture of the state of the research on several
topics related to distribution system networks; 2) collecting and making accessible some
technical indicators regarding the main DSOs in Europe, and 3) building representative
distribution networks that could be used to perform different kinds of analysis and
assessments without the need to have access to the real DSOs data.

During the study, the JRC collaborated with the IIT of Comillas Pontifical University,
which supported the project with its established expertise and knowledge in the analysis
and development of models for the simulation and optimization of future electricity
networks.

One of the main challenges of the project was collecting enough data to substantiate the
analysis. An online survey was launched in January 2015 (

12
Annex C: On-line survey) with the aim of collecting several clusters of data, relating to
types of ownership and unbundling, network structures and designs, amounts and types
of connected distributed generation (DG), and reliability of supply indicators. The survey
was available on the EUSurvey web platform, where DSOs could directly fill in their data.
In order to reach out to the DSOs and speed up the data gathering activity, the JRC
teamed up with EURELECTRIC, an electricity industry sector association. Thanks to its
collaboration, a link to the online questionnaire was sent to all the European DSOs that
are subject to the EU unbundling requirements, i.e. those serving more than 100,000
connected customers.
The questionnaire was divided in two parts. The first part was mandatory and it was
aimed at collecting general company information and some basic parameters on their
networks design, e.g. the number of customers and circuit length per voltage level or
the total installed capacity of generation connected. The response rate was quite high,
as 79 out of the 190 larger DSOs replied.

The second part of the survey asked for additional information relating to the network
structure, reliability indexes and connected distributed generation. Filling in this part of
the questionnaire was only optional, but several DSOs decided to share their data: 20
DSOs provided information on their network design structure, 22 on the generation
connected to their distribution network and 18 on the reliability indexes for long
unplanned interruptions. In addition, 40 DSOs gave their availability for providing more
customized information with the purpose of building representative networks of their
company/country.

On the basis of the information collected through the survey, several network structural
indicators were constructed, linking DSOs' inputs (e.g. the number of connected
consumers, supplied area) and outputs (e.g. circuit length per voltage level, number and
capacity of HV/MV and MV/LV substations). Chapter 3 presents and details the main
indicators, highlighting the differences between DSOs and trying to account for the main
reasons behind these differences.

Building on the structural indicators, 13 representative networks were constructed: 3

large-scale and 10 feeder-type networks. These grids are representative of European
networks, but do not specifically represent any particular DSO or country. Chapter 4
introduces the concept of reference network models and guides the reader through the
process for their construction.

Finally, Chapter 5 presents different simulation analyses that have been carried out to
illustrate the potential application of these representative networks. One example of
such simulations is the study of the impact of growing DG penetration on voltages and
thermal limits. This type of analysis aims at identifying solutions to maximize renewable
penetration with no need of additional network investment and can therefore be very
useful for technical and regulatory purposes.
Overall, the DSO Observatory project successfully addressed all its main objectives. This
study represents a first contribution to better understand and address the challenges
DSOs will have to face in the transition to the new energy system. The JRC will continue
its scientific and policy support activities in the field; this report can therefore be seen as
a first step of an exercise that will continue in the future.

13
3 A clearer view of the European distribution networks

This chapter presents the results of the survey and explains how the collected data were
used to build the indicators. From the list of 36 indicators derived by relating DSOs input
and output data, 10 indicators were selected as the most relevant to build the
representative distribution networks. Each indicator of this smaller subset is presented
and explained in the following paragraphs, highlighting the differences between DSOs
and trying to account for the main reasons behind these differences. To protect the
confidentiality of the collected data, DSOs were anonymized. In each graph, they are not
referred to by their name or by the country where they operate but they are numbered
according to their position on the X-axis, which varies for each indicator. At a later
stage, and subject to the conclusion of ad-hoc agreements with the DSOs, data by
country will be published online on a dedicated JRC webpage3.

3.1 Survey participation

79 out of the 190 larger DSOs responded to our survey. Even if the response rate was
lower than 50%, the representativeness of the sample is quite high. Together, the 79
DSOs distribute more than 2,000 TWh of electricity to over 200 million customers per
year, covering a total area of more than 3 million square km. They distribute over 70%
of the electricity distributed by all DSO serving over 100,000 customers (Eurelectric
2013).

To provide a general perspective of the DSOs sample, two aggregated values are shown.
Figure 3-1 shows the distribution of connected customers served by the DSOs in the
sample. More than half of them serve between 400,000 and 4 millions of customers.
Figure 3-2 shows the yearly distributed energy. In this case, more than 60% of the
DSOs distribute between 3 and 60 TWh (of energy) per year to their customers.

3
More information will be published on http://ses.jrc.ec.europa.eu

15
 40%

35%

30%
Percentage of DSOs 25%

20%

15%

10%

5%

0%

Number of connected customers

Figure 3-1 Distribution of connected customers

40%

35%

30%
Percentage of DSOs

25%

20%

15%

10%

5%

0%

Average distributed energy (GWh)

Figure 3-2 Distribution of yearly distributed energy

Table 3-1 shows for each country the number of respondents with respect to the total
number of DSOs serving more than 100,000 customers (first column) and the
percentage of connected customers over the total number of customers connected by
the larger DSOs (second column).

16
Together, the 79 DSOs cover 74.8% of the total number of customers connected to the
larger DSOs. In most countries this figure is over 70%, while in a few countries (e.g.
Italy, Check Republic), it reaches 100%. Yet, there are a few cases with very small or
even zero coverage (e.g. Austria, Malta). All together however, the 79 DSOs can be
considered as a good representation of the investigated DSOs population.

Customers
Country No. of DSOs
covered
Austria 2/13 6.5%
Belgium 2/15 77.0%
Bulgaria 1/3 34.3%
Croatia 1/1 100.0%
Cyprus 1/1 100.0%
Czech Republic 3/3 100.0%
Denmark 3/6 41.2%
Estonia 1/1 100.0%
Finland 2/7 23.0%
France 1/5 96.0%
Germany 28/75 49.3%
Greece 1/1 100.0%
Hungary 3/6 45.9%
Ireland 1/1 100.0%
Italy 3/3 100.0%
Latvia 1/1 100.0%
Lithuania 1/1 100.0%
Luxembourg 1/1 100.0%
Malta 0/1 0.0%
Netherlands 2/8 39.0%
Poland 5/5 100.0%
Portugal 1/3 99.0%
Romania 2/8 78.5%
Slovakia 2/3 71.2%
Slovenia 1/1 100.0%
Spain 3/5 95.3%
Sweden 3/6 37.3%
United Kingdom 4/7 44.3%
Total general 79/190 74.8%

Table 3-1 Participation per country

Figure 3-3 visualises the same information through a colour-coded map. The different
shades of red show how the percentage of covered customers varies across the EU 28,
with the darker shades representing a higher coverage.

17
 Figure 3-3 Customers coverage per country

3.2 DSOs indicators

This section presents 10 of the 36 indicators (Table 3-2, Table 3-3 and Table 3-4) that
we have built from the collected data provided by the 79 DSOs in the database. The
indicators have been devised in such a way to allow for a comparative analysis of the
DSOs and can be divided in three main categories:

 Network structure
 Network design
 Distributed generation

18
 Network structure and reliability indicators
1. Metrics associated to LV network
LV consumers per area
LV circuit length per LV consumer
LV circuit length per area of distribution
LV underground ratio
2. Metrics associated to MV/LV substations
Number of LV consumers per MV/LV substation
Area per MV/LV substation
Capacity of MV/LV substations per consumer
Area covered per capacity of MV/LV substation
3. Metrics associated to MV network
Number of MV consumers per area
MV circuit length per MV supply point
MV circuit length per area of distribution
MV underground ratio
4. Metrics associated to HV/MV substations
Number of MV supply points per HV/MV substation
Area per HV/MV substation
Capacity of HV/MV substation per MV supply point
Ratio of capacity of MV/LV substations per capacity of HV/MV substation
Area per capacity of HV/MV substations
5. Metrics associated to HV network
HV circuit length per HV supply point
HV circuit length per area
HV underground ratio
6. Other relevant metrics
Number of electric vehicle public charging points per consumer
SAIDI for long unplanned interruptions
SAIFI for long unplanned interruptions

Table 3-2 Network structure and reliability indicators

The network structure (Table 3-2) data covers DSO inputs referring to the main
parameters (such as the number of connected consumers, distributed generation, area
of supply and distributed annual energy) and DSO outputs corresponding to the assets
planned by the DSO to cope with the given inputs. The outputs are divided by voltage
level (LV, MV and HV) and consist of circuit length, number and capacity of substations,
etc.

Given the structure of the data one can relate inputs and outputs through ratios in the
following way: Input/Input (I-I), Input/Output (I-O), Output/Input (O-I) and
Output/Output (O-O). The number of consumers and the full covered area are for
instance the most relevant inputs to build the network structure indicators. Mainly
invariant ratios O-I or I-O are calculated, e.g. LV circuit length per LV consumer or area
covered per HV/MV substation. This type of indicators relate the inputs, which are the
structure of the demand or the DG that the DSO must connect, with the outputs, which
consist of the installations that the DSO uses to cover that given demand and to connect
that given DG.

19
Additionally, network structure indicators of O-O or I-I type are calculated. The O-O
indicators (such as capacity of MV/LV substations per capacity of HV/MV substation),
analyze design criteria used by DSOs when sizing their network installations. The I-I
indicators (such as consumers per area) analyze the structure of the demand and DG
that the DSOs have to connect to their networks. Note that the DSOs have no control on
the I-I indicators, while their planning decisions can affect the rest of indicators.

The network design indicators (Table 3-3) comprehend metrics associated to

substations and feeders, and other relevant metrics. This information aims to identify
which are the typical parameters that are used by DSOs for sizing and designing
distribution installations.

Network design indicators

1. Metrics associated to substations

Typical transformation capacity of HV/MV substations

Typical transformation capacity of the MV/LV Substations in urban areas
Typical transformation capacity of the MV/LV Substations in rural areas
Average number of MV/LV substations per feeder in urban areas
Average number of MV/LV substations per feeder in rural areas
2. Metrics associated to feeders

Average length per MV feeder in urban areas

Average length per MV feeder in rural areas
3. Other relevant metrics

Voltage levels
Automation equipment and degree of automation

Table 3-3 Network design indicators

The analyses include for instance the typical transformation capacities of the HV/MV and
MV/LV substations, how many MV/LV substations are connected to a MV feeder, and the
average length of MV feeders. This information is generally presented broken down in
urban and rural areas, as the network design depends on the type of area (e.g. MV
feeders use to be longer in rural areas than in urban areas).

Distributed generation indicators

Total Installed Capacity/consumer
Percentage of generation connected to LV per technology
Percentage of generation connected to MV per technology
Percentage of generation connected to HV per technology

Table 3-4 Distributed generation indicators

The distributed generation indicators (Table 3-4) for the analysis of DG refer to the
installed capacity per consumer, and the percentage of DG connected to each voltage
level. Typically, the total installed capacity is available in the literature but it is not
usually broken down per voltage level.

20
3.2.1 Main DSOs indicators explained
In the following a subset of the indicators (listed in Table 3-5) used to build the large-
scale representative distribution networks is shown. In these graphs the DSOs are sorted
by the indicator value, moving from the lowest to the highest value. This solution has
been adopted mainly to protect the confidentiality of the data and the anonymity of the
participants in the survey, as requested by certain DSOs. This means for instance that
the DSO in position #1 in Figure 3-4 is the DSO with the lowest number of LV consumers
per MV consumer, and it is not necessarily the same DSO in position #1 in Figure 3-5,
which is instead the DSO with the lowest LV circuit length per LV consumer.

ID DSOs indicators

1 Number of LV consumers per MV consumers

2 LV circuit length per LV consumer

3 LV underground ratio

4 Number of LV consumer per MV/LV substation

5 MV/LV substation capacity per LV consumer

6 MV circuit length per MV supply point

7 MV underground ratio

8 Number of MV supply points per HV/MV substation

9 Typical transformation capacity of MV/LV secondary substations in urban areas

10 Typical transformation capacity of MV/LV secondary substations in rural areas

Table 3-5 Subset of the total DSOs indicators used to build the large-scale representative
distribution networks

The number of LV consumers per MV consumer is a measure of the percentage of LV

residential & commercial and MV industrial consumers. The median for this indicator
corresponds to 354 LV consumers per each MV consumer. From the figure 3-5 a
noticeable gap is easily identified with a minimum of 10 and a maximum of 2008 LV
consumers per MV consumers. This large gap is due to the fact that certain DSOs serve
very large populated areas (urban areas) while others are serving more rural areas.

21
 Figure 3-4 LV consumers per MV consumer

The LV circuit length depends mainly on the location of LV consumers and the
distances among them. The median LV circuit length per LV consumer is around 0.023
km/LV consumer, with a maximum of 0.078 km/LV consumer and a minimum of 0.008
km/LV consumer. The LV circuit length per LV consumer is higher in more rural countries
than in more urban countries because typically in cities, electric density is higher and the
LV feeders have to be shorter. However, this also depends on the voltage levels of the
distribution networks.

Figure 3-5 LV circuit length per LV consumer

The LV underground ratio is defined as the length of LV underground circuits divided

by the total length of LV circuits (overhead and underground). The median LV
underground ratio is 78%, with a maximum of 100% and a minimum of 11.5%. It is
argued that a high underground ratio is desirable to reduce the number of interruptions

22
as the failure rate of this type of conductors is lower. However, in practice the decision of
undergrounding cables is generally taken for aesthetic criteria (e.g. people do not want
to have overhead electrical lines near their houses), rather than based on technical
criteria. Typically the underground ratio is higher inside settlements than outside
settlements. Urban areas with many consumers usually have more underground cables.

Figure 3-6 LV underground ratio

The median number of LV consumers per MV/LV substation is around 90, with a
maximum of 278. This ratio is interesting because it gives an idea of the size of the low
voltage network below each MV/LV substation. Of course, the length of the network will
depend also on the dispersion of the consumers, being different in rural and urban areas.
In urban areas, in which the capacity constraint is the most relevant one, the main
parameters which affect this ratio are the peak power of the LV consumers and the
capacity of the MV/LV substations. However in rural areas, voltage constraints can be
also very relevant, and in that case this ratio also depends on the LV circuit lengths.

23
 Figure 3-7 Number of LV consumers per MV/LV substation

The capacity of MV/LV substation per LV consumer is an indication on how much

power is installed in the MV/LV substation for each LV consumer. This basically depends
on the average typical peak power of the consumers, and on simultaneity factors. The
higher the peak power of consumers, the higher the capacity of MV/LV substation
required. In distribution areas in which the MV/LV substations are oversized, this value
can also be higher.

Figure 3-8 Transformers capacity per LV consumer

The median capacity of MV/LV substations per LV consumer is around 3.66 kVA, with a
maximum of 11.4 kVA and a minimum of 2.1 kVA. This ratio depends, among others, on

24
the type of the household. For instance, a bungalow and a big apartment hosting more
than four people have different load profile.

The MV supply points are MV/LV Substations and MV consumers. These are the key
items of interest for the MV network when MV DG 4 is not connected or available. For
evaluating the MV network length, the ratio MV circuit length per MV supply point
was selected. As the number of MV supply points is much lower than the number of LV
consumers, this parameter is much higher than the LV circuit length per LV consumer.
The median MV circuit length per MV supply point is in fact around 0.73 km/(MV Supply
Point), with a maximum of 1.86 km/(MV Supply Point) and a minimum of 0.11 km/(MV
Supply Point). Despite the fact that there is so much variability in the number of MV
consumers per area, the MV circuit length per MV supply point does not differ so much
among different regions, meaning that the DSO can have some control on this variable,
but the range is not so wide.

Figure 3-9 MV circuit length per MV supply point

The MV underground ratio is defined as the length of MV underground circuits divided

by the total length of MV circuits (overhead and underground). In general the
underground ratio is lower in rural areas than in urban areas. The underground ratio can
have a side effect on reliability, as underground cables usually have lower failure rates
(improving reliability) but requiring higher repair times. However, underground cables
usually require more investments, due especially to the high cost of making the ditches
rather than those related to the cost of cables themselves. Nevertheless sometimes the
decision of installing underground or overhead cables could depend on aesthetic issues.

4
MV DG was not considered due to the fact that the number of DG installations connected to each voltage
level was not available.

25
 Figure 3-10 MV underground ratio

The median MV underground ratio is around 56.8%, with a maximum of 100% and a
minimum of 9.4%. The median LV underground ratio was 78%, higher than in MV,
possibly due to lengthy overhead MV feeders connecting several settlements. The large
distance between the maximum and the minimum values could suggest that the DSOs
have the freedom to choose this ratio; however specific national or regional regulations
can impose a mandatory solution, meaning that the DSO cannot optimize this parameter
in its planning.
The HV/MV substations have to supply electricity to MV supply points (MV consumers
and MV/LV substations), apart from connecting MV distributed generation if any. The MV
consumers and MV/LV substations are distributed along feeders, and therefore the
number of MV supply points per HV/MV substation is the product of the number of
feeders of the substations and the average number of MV supply points per feeder. Both
of them are ratios which depend on the structure of the distribution networks, having the
DSOs some control on them. The median number of MV supply points per HV/MV
substation is 178, with a maximum of 1210.

26
 Figure 3-11 Number of MV supply point per HV/MV substation

The transformation capacity of MV/LV substations in urban areas (Figure 3-12) is

typically higher than in rural areas, due to the increased electricity density. The most
typical value of MV/LV substation capacity in urban areas is 630 kVA. The median
transformation capacity of MV/LV substations in urban areas is 630 kVA, with a
maximum of 1000 kVA and a minimum of 400 kVA.

The transformation capacity of MV/LV substations in rural areas (Figure 3-13) is

generally lower than in urban areas, due to the reduced electricity density and increased
distances. The most typical value of MV/LV substation capacity in rural areas is 400 kVA,
followed by 250 kVA and 100 kVA. The maximum transformation capacity is 630 kVA
and the minimum is 50 kVA.

Figure 3-12 Typical transformation capacity of the MV/LV secondary substations in urban areas
(kVA)

27
 Figure 3-13 Typical transformation capacity of the MV/LV Secondary Substations in rural areas

The remaining indicators listed in Table 3-3, Table 3-4 and Table 3-5 are presented in
the Annex B: Other indicators.

28
4 From Distribution Systems to Representative
Distribution Networks

In this chapter a direct connection is established between the indicators presented in the
previous chapter and the building process of representative distribution networks using
the Reference Network Model. After introducing the concept of reference networks and
their usage in literature, the methodology used in (Mateo, et al. 2011) is explained. Then
the 13 typologies of networks (3 large-scale and 10 feeder-type topologies) that have
been built through the collected data provided by the DSOs are described in detail.

4.1 The Reference Network Model concept

After the restructuring of the electric power sector, which started in many countries
during the 1990s, electricity transmission and distribution being regulated activities
started to be remunerated through new regulatory approaches. One of the key means
used by energy regulators to apply these new regulatory tools was the concept of
Reference Network Models (RNM) to build a reference utility (Rudnick 2000) (Román
1999). RNM are very useful tools also when several scenarios for the growth of real
networks needs to be taken into account. Network expansion, reinforcements and
maintenance costs necessary to accommodate potential increases in loads and
penetration of distributed energy sources can be assessed (Jenkins 2014). Additionally,
when performing costs-benefits analysis5 (CBA) of smart grids projects, RNMs can also
be used when future scenarios need to be considered.

There are two common techniques to build representative networks. One approach starts
with the DSOs real networks and by applying clustering techniques, it tries to identify
the most typical configurations of the considered networks. When real networks are not
available, an alternative option is to build synthetic networks, using a large-scale
distribution network planning tool. As real networks were not available in this study, the
second approach has been used. Note that this approach coincides with that of a DSO
whose aim is to distribute electricity in a given area of interest fulfilling present technical
and physical constraints.
A RNM is a large-scale network distribution planning tool. Extensive data are required
and several algorithms are run to obtain the design of a distribution network
representation which can be very close to the real one. The RNM designs the distribution
grid in a considered area of service from the transmission substations to the final
consumers considering three hierarchical levels: high voltage, medium voltage, and low
voltage. Both rural and urban zones included in the considered area of service are
simultaneously optimized by designing the grid to supply all the customers (from
thousands to millions) and including the connected distributed generators. Geographical
constraints regarding streets topologies in urban areas and environmental factors
(coastline, natural reserves, mountains, etc.) can be taken into account in the design.

5
A CBA methodology for Smart Grid Projects is reported in (V. Giordano 2012).

29
The continuous interaction between the electrical network design algorithms and the
Geographical Information System (GIS) ensures the feasibility and optimality of the
layout and location of the planned installations (Gómez 2013).
From a computational point of view, the network planning is a difficult optimization
problem whose complexity increases with the increasing number of consumers to be
modeled (Mori 2003). To overcome these issues, RNMs address the problem through the
use of heuristics, and considering several distribution areas, so that each of them can be
planned separately, in order to diminish computing times by paralleling processes.

Two types of RNMs can be developed:

 A green-field model which designs the whole distribution grid, including
substations and power lines, from scratch;
 An expansion model which designs the reinforcement and new additions to the
existing network needed to face future situations, for instance growth of demand
or new distributed generation.

In both models, the design problem consists in minimizing the total investment in new
installations plus associated operational costs, mainly energy losses, in order to supply
the expected demand while meeting reliability and quality of supply criteria (e.g. SAIDI,
SAIFI, voltage quality, etc.). In this study only models of the first type (green-field) have
been built. In the following three sections a technical explanation of the obtained outputs
and the steps to follow to obtain them from the required inputs is given. The reader not
interested in this level of detail can move directly to section 4.1.4 where the general
methodology to connect the indicators to the representative networks is explained.

4.1.1 RNM Outputs

The result obtained by the RNMs is composed of two main layers. One provides a
summary of the most relevant information of the designed network and the
corresponding costs adequately broken down per type of network component.
Additionally, the reliability indices obtained for the given network are provided.

The second one provides a collection of graphical files created by the RNMs. Each of
these files corresponds to a type of network component including not only geographical
information (GIS), but also electrical information such as impedances, thermal capacity
and peak power flow.

4.1.2 RNM Inputs

To obtain the output previously explained, an extensive amount of data is required for
building the RNM. As expected, the finer the available data the more accurate the built
representative network can be.
The required input data can be grouped in four main categories: network users,
transmission substations, network components and other relevant parameters (Gómez
2013).
“Network users” refers mainly to the identification and description of loads, Distributed
Generation (DG), Electric Vehicles and storage units connected to the distribution
network. The exact location of every single user (X and Y coordinates), the voltage level

30
at the point of connection, the contracted power or the installed capacity and power
factor are key data for the RNM. For large distribution areas, the number of consumers,
particularly small LV consumers, can be very large. The networks studied can range from
several thousand users up to a few million connections. Gathering such very detailed
information is one of the main difficulties of the RNMs.
“Network components” indicates a library of standard components that includes
technical data on HV power lines, MV and LV feeders, HV/MV substations, MV/LV
transformers, protection equipment (breakers, fault detectors and switches),
maintenance crews, capacitor banks and voltage regulators. For each network
component in the library, investment and maintenance costs, rated capacity, electrical
properties such as impedances, and useful life of the component are specified. To
compute the expected reliability indices it is also necessary to provide the models with
failure rates and a standard annual duration of preventive maintenance actions that are
carried out on each type of component (HV line, MV/LV transformer, etc.), for overhead
and underground elements and in each kind of area (urban or non-urban).
“Others relevant parameters” refers to the remaining parameters which the RNM
needs in order to perform a reliable computation. The most relevant ones are
summarized as follows:

Simultaneity factors. Simultaneity factors are needed for planning purposes, in order to
take into account that the maximum power flow in the different network components
does not occur at the same moment. As the grid voltage level rises, more downstream
customers and installations are aggregated. However, the peak of an upstream network
element is lower than the sum of the peaks of its downstream fed network components,
because they do not all occur at the same time. Therefore a simultaneity factor has to be
considered when peak power flows are aggregated. Without simultaneity factors, LV
grids and MV/LV transformers would be much bigger in terms of capacity than what it
would be actually required. Similarly, MV/LV transformers and distribution substations
have two different simultaneity factors, one upstream of the transformer and another
downstream. The upstream simultaneity factor takes into account that not all
transformers are at their peak at the same moment. The downstream simultaneity factor
models the fact that not all the lines connected to the transformers will be loaded at
their maximum simultaneously.

Economic parameters. Economic parameters are needed to calculate the present value of
network costs and evaluate investment options. These comprise the cost of energy
losses, the weighted average cost of capital (WACC), and the costs to install conductors
in different types of areas.

Load modelling and GIS related parameters. Density and minimum number of
consumers to classify them into different areas and identify settlements, degree of
undergrounding required within settlements per voltage level, and street maps
parameters.

Technical and quality constraints. The RNMs must observe the maximum and minimum
bus voltages, the thermal constraints and the limits imposed on reliability of supply
indices. The RNMs use the SAIDI (System Average Interruption Duration Index) which is
a measure of the duration of the interruptions and the SAIFI (System Average
Interruption Frequency Index) which measures the frequency of the interruptions. Bus
voltage limits are set per voltage level and should comply with the limits imposed in the
EN 50160. MV network must comply with zonal and individual reliability indices, which

31
are separately fixed for urban, semi-urban, concentrated rural, scattered rural and
industrial areas.

4.1.3 RNM planning steps

One of the first steps for building a RNM is to decide the number, size and location of
MV/LV power transformers once the inputs on LV final customers and distributed
generation are given. An initial estimation of the number of MV/LV transformers is made
on the basis of the power density of each identified city or town. An algorithm is used to
locate the proposed MV/LV transformers.

After locating MV/LV transformers, the LV network can be planned connecting the LV
final customers and the distributed generation. The process is carried out as follows.
First, the Delaunay algorithm is run with all input data. Second, each final customer and
distributed generator is associated to a MV/LV transformer, applying an electric
momentum criterion, to obtain several clusters with a unique MV/LV transformer. Third,
a minimum spanning tree is run in each cluster, where MV/LV transformers are the root
nodes of the tree. Fourth, a branch-exchange optimization algorithm is executed to
estimate a quasi-optimum LV network, subject to voltage and current constraints,
minimizing investment, and operation and maintenance costs. This optimization
sometimes implies relocating the LV/MV transformers and thus, returning to the first
step. Finally, the conductor size optimization is performed to select the optimum for each
LV overhead or underground line section. To this end, an additional term that takes
losses into account is incorporated into the objective function.

The third stage consists in deciding the number, size and location of HV/MV substations.
The logic for algorithms at this stage is similar to the one described for LV/MV
transformers planning.

The fourth step of the algorithm is to map out the MV network, which will link MV/LV
transformers to HV/MV substations. The process resembles LV network planning, despite
including new features that allow taking the quality of service level into account.

4.1.4 RNM Methodology

In the following, the methodology used to build European large-scale representative
networks is presented. The typical topologies of EU distribution networks have been
modeled. The voltage levels considered are low voltage (LV) and medium voltage (MV).

The constructed representative networks can be categorized into two major groups:

 Large-scale networks (cases #1-3) model the network downstream of a HV/MV

substation, including LV & MV consumers, LV & MV feeders and MV/LV
substations. The main difference between these large-scale networks is the
density of the demand. An urban network has been built modelling the
distribution network inside a highly populated city. A rural network has also been
built modeling farms and small settlements connected by a MV network. Finally a
semi-urban network represents an intermediate situation in the outskirts of a
city.

32
  Feeder type networks (cases #4-13). These networks include feeders
downstream a single HV/MV or MV/LV substation and can be divided in:

a. MV feeders (cases #4-11), with specific network configurations for MV

reliability analysis, aimed at assessing the duration and frequency of the
interruptions depending on the network configuration and the MV
protection equipment. Networks with a low and high degree of automation
have been produced in this case to be able to evaluate how the degree of
automation impacts on the continuity of supply. Networks with a low
degree of automation use fault detectors, switches and breakers which are
manually operated; while networks with a high degree of automation use
tele-controlled versions of some of these devices. In networks with a high
degree of automation, typically tele-controlled devices are regularly
spaced along the feeders to maximize reliability improvement.

b. LV feeders (cases #12-13), aimed at making analysis of a single LV

network, for example for installing photovoltaic generation in a low voltage
residential network. Each of them represents the network downstream a
MV/LV substation. Network #12 models an urban LV network while
network #13 models a semi-urban LV network.

Table 4-1 summarizes the 13 representative networks built.

Representative Voltage Degree of

Type of area
network ID # levels automation

1 Urban LV & MV Low

2 Semi-urban LV & MV Low
3 Rural LV & MV Low
Urban - Two substations
4 MV Low
interconnected
Urban - Two substations
5 MV High
interconnected
Urban - One substation and one
6 MV Low
switching station
Urban - One substation and one
7 MV High
switching station
8 Semi-urban - Substation ring MV Low
9 Semi-urban - Substation ring MV High
10 Rural MV Low
11 Rural MV High
12 Urban LV Low
13 Semi-urban LV Low

Table 4-1 Representative networks

The selected topologies and particular networks represent typical cases of EU distribution
networks but other alternatives and variants could also be considered for specific
analysis depending on their objective and scope.

33
Each synthetic network has been obtained using a Greenfield Reference Network Model
(RNM). As already mentioned, the RNM requires as input the location and peak demand
of every single customer. The geographical coordinates of the consumers for each type
of network (urban, semi-urban, and rural) were determined by processing a street map
image, downloaded from OpenStreetMap 6 . Next, this information was complemented
with the peak demand of each consumer and a RNM catalogue (containing among others
standard network installations). As a result, the synthetic network was built. Finally, an
iterative procedure was carried out by readjusting the inputs of the model until the
parameters of the obtained synthetic networks matched the network structural indicators
of DSOs in the European Union (presented in section 3.2).

When adjusting the RNM input parameters, the building density and the number of
consumers per building can help to increase or decrease the consumer density, which in
the end modifies the network length per supply point. The percentage of MV consumers
with respect to the total number of consumers is both a RNM input parameter and a
structural indicator, so it can be directly adjusted to the desired value. Since in the
synthetic network there is only one HV/MV substation the size of the covered area
determines the number of demand points supplied by the substation and this needs to
coincide with the structural indicator (number of MV supply points per HV/MV
substation). The distribution function of the consumer peak demands and the
simultaneity factors impact on the two following structural indicators: i) required
installed capacity of MV/LV substation per LV consumer, and ii) the number of LV
consumers per MV/LV substation, so they have been adjusted iteratively. The network
underground ratios per voltage level inside settlements are RNM input parameters that
can help to adjust the total underground ratios. Finally, the MV/LV transformer capacities
in the RNM installation library can be directly adjusted to the typical values obtained for
the EU DSO structural indicators, which are broken down in rural and urban areas.

Figure 4-1 shows the different steps followed by the proposed methodology to construct
synthetic representative distribution networks.

Large-scale networks represent the grid downstream of a single HV/MV substation,

including LV & MV consumers, LV & MV feeders and MV/LV substations. Three grids have
been built corresponding to urban, semi-urban and rural areas. The label of large-scale
is due to the high number of buses present in each network of this kind. They are geo-
referenced, because the topological coordinates of each single bus are provided.

1. The first step is to take the European DSO network structural indicators from the
DSOs database and to use them as a reference to validate the obtained EU
synthetic network. These indicators represent those network parameters that the
built distribution network needs to have in order to be considered a
representative proxy of a European network.
2. The next step for each representative network is to obtain a street map image,
which will help to identify and select the location of buildings. Different images
were considered for each type of area (urban, semi-urban or rural), trying to
represent the typical consumer density in these types of locations. For example,
in the urban network the centre of a city was selected, while in the rural network
a countryside region with small settlements and farms was chosen. To gather the
geographical coordinates (X,Y) of the buildings the street map image has to be

6
https://www.openstreetmap.org/

34
 processed, setting a density of buildings inside blocks (which determines the
spacing between buildings) and the size of the area (which determines in the end
the number of consumers supplied by the HV/MV substation).
3. Then the main parameters characterizing the consumers need to be defined:
voltage level (LV or MV), geographical position coordinates (X, Y) and peak
demand. In order to obtain these parameters, the percentage of MV consumers in
the supplied area, the number of consumers per building, and an estimate of the
distribution function of the consumer peak demands are specified. First, the
percentage of MV consumers is used to fix the proportion of LV & MV consumers,
determining their voltage levels at the grid connection points. This ratio was
obtained as a structural indicator. Then, the number of consumers per building is
used to place several consumers in each building, locating them in the same
geographical coordinates. Therefore consumer coordinates are based on the
coordinates of the buildings collected through the street map image processing.
Finally, the distribution function of the consumer peak demands is used to set the
peak demand of each particular consumer depending on its voltage level. The
distribution function of the consumer peak demands and the number of
consumers per building are adjusted at each iteration of the proposed
methodology, see Figure 4-1.
4. A RNM input file is also provided, defining simultaneity factors, settlement
underground ratios, and identifying the standard network installations that the
RNM will use to build the synthetic networks.
5. Once all the aforementioned data are gathered, the Greenfield Reference Network
Model (RNM) is run to obtain a first proxy synthetic network. Such a network is
mainly dependent on the pre-specified consumer data, location of the HV/MV
substation and the standard installation library included in the model. The
Greenfield RNM uses as the main control variables: simultaneity factors,
settlement underground ratios, and the catalogue of standard network
installations.
6. Finally the resulting structural indicators that correspond to the obtained RNM
synthetic network are measured and compared to the EU DSO structural
indicators, which are used as reference. If the two sets of indicators are
reasonably close the synthetic network can be chosen as a representative proxy
of the target network.
7. In practice, it is always necessary to iterate this process several times,
readjusting the RNM input parameters until the convergence to the target
structural indicators is achieved.

35
Figure 4-1 Schematic view of the methodology used to build the representative distribution
networks

36
4.2 Realistic synthetic networks based on RNM
To build the synthetic networks using the RNM, the location of the customers have to be
identified. For this purpose street map images need to be downloaded and processed.
For the urban area, the center of a city was selected. In the rural area a region with
small settlements and farms was chosen. Finally, the semi-urban area was taken as an
intermediate one, representing the outskirts of a city.

The following figures show how the street map images were processed to obtain the
location of consumers. Starting with a real street map image, in the urban area first the
streets of the city were identified as shown in Figure 4-2.

Figure 4-2 Street map image processing: identification of streets in the urban network

Then, the block of buildings was recognized, and finally the location of consumers was
selected placing them around blocks of buildings as shown in Figure 4-3 where the blue
points represent the buildings. All the consumers inside a building were placed in the
same coordinates, as they correspond to a single bus from the network point of view.

Figure 4-3 Street map image processing: building location in the urban network

37
A similar process was carried out in the semi-urban and rural areas, but directly
identifying blocks of buildings, instead of starting with the streets. As shown in Figure
4-4 the outskirts of a city were modeled in the semi-urban area.

Figure 4-4 Street map image processing: semi-urban network

As shown in Figure 4-5, the rural area selected includes small settlements and farms
around the settlements.

Figure 4-5 Street map image processing: rural network

Once the location of consumers was obtained, this information was complemented with
the peak demand of each consumer, and with the RNM input catalogue. All this
information was then used as input to the RNM in order to obtain the synthetic networks
of each area.

Table 4-2 shows the median, the 0.05 and the 0.95 percentile of the values collected in
the DSO Observatory database. The 0.05 and 0.95 percentiles were chosen with the aim
of covering the full diversity of DSOs, but removing outliers. The values that were used
as reference for the RNM to build the networks are marked in light blue. In general the
median is the more relevant value, however for underground ratios, a great diversity
exists, and therefore in this case the three values have been considered.

38
 0.05 0.95
Network indicators Median
percentile percentile

Number of LV consumers per MV consumer 26 354 1550

LV circuit length per LV consumer (km) 0.0104 0.023 0.0542

LV underground ratio 22.20% 79% 99.80%

Number of LV consumers per MV/LV substation 24.9 90 209.2

MV/LV substation capacity per LV consumer (kVA) 2.37 3.8 9.12

MV circuit length per MV Supply Point (km) 0.27 0.73 1.46

MV underground ratio 14.40% 56% 100%

Number of MV Supply Points per HV/MV

61 178 580
substation

400-1000 Urban
MV/LV transformer substation capacity (kVA)
100-400 Rural

Table 4-2 DSO Database network indicators

4.2.1 Representative large-scale networks

Table 4-3 shows the calculated indicators that correspond to the constructed
representative networks (Urban, Semi-Urban, Rural). The first column is the indicator
name the next three columns are the values of each indicator for the constructed
representative networks. It is worth noticing that the European DSO indicators in the
database are not broken down in urban, semi-urban and rural areas because this
information was not provided with that level of detail by the DSOs, with the exception of
the typical MV/LV substation capacities. This implies that some variations exist between
representative network (Table 4-3) and target (Table 4-2) indicators, for example, as it
is the case of purely urban networks.

Network indicator Urban Semi-urban Rural

Number of LV consumers per MV consumer 335 350 386

LV circuit length per LV consumer (km) 0.004 0.008 0.027

LV underground ratio 86% 42% 4%

Number of LV consumers per MV/LV

101 87 51
substation

MV/LV substation capacity per LV consumer

6.499 6.995 5.209
(kVA)

MV circuit length per MV Supply Point (km) 0.2 0.3 0.8

MV underground ratio 100% 74% 15%

Number of MV Supply Points per HV/MV

164 201 172
substation

MV/LV transformer substation capacity 400 630 100 250 400 630
100 250 400
(kVA) 1000 1000

Table 4-3 Indicator name and representative network ratios

39
Annex A: Indicator box plots, reports the box plots of the first eight indicators listed in
Table 4-3. These box plots provide a graphical benchmarking between the values
obtained through the DSOs data and the values measured from the representative
networks that have been built through the methodology explained in section 4.1.4.
The following sections show the graphical representation and the main data of the three
large-scale geo-referenced networks obtained through the RNM. The choice of using the
following such networks depends much on the different possible configurations,
(topologies, voltage levels, number of consumers, etc.) which may be used in real
situations.

4.2.1.1 Urban (#1)

Figure 4-6 shows the large-scale network in an urban settlement, including MV feeders
(blue lines), LV feeders (black thin lines), as well as MV/LV substations (red circles) and
the HV/MV substation (blue triangle). All MV feeders are underground.

Figure 4-6 Urban network

Aggregated data are provided in the following to give an idea of the size of the
considered distribution network and of its main characteristics. All the representative
distribution networks built within the DSOs Observatory project will be publicly
downloadable on the SESI webpage7.

7
http://ses.jrc.ec.europa.eu/distribution-system-operators-observatory

40
The number of MV consumers is much lower than the number of LV consumers, in the
proportion showed by the reference indicators.

Aggregated inputs
No. Peak Power (MW)

LV Consumers 12735 56.78

MV Consumers 38 5.7

Since we are modeling a city centre, no MV overhead conductors are present.

Overhead Underground
LV km 7.38 46.49
MV km 0 31.20

Even though there is one single HV/MV substation and 126 MV/LV substations, the
relation between the total capacity of MV/LV substations and HV/MV substations is close
to one, as the reference indicators shows (please see Figure B-9).

No. Capacity (MVA)

MV/LV substation 126 82.76
HV/MV substation 1 80

In the following table the label “type ID” indicates the type of electrical line, according to
the labels specified in section 4.2.2.6. Types LV_UO_1/2 refer to overhead and types
LV_UU_1/2 to underground. Most LV feeders are underground, but there are also a few
LV overhead feeders, corresponding to electrical lines in façades.

Low voltage network

Voltage (kV) Type ID Length (km)
0.4 LV_UO_1 2.929
0.4 LV_UO_2 4.45
0.4 LV_UU_1 38.956
0.4 LV_UU_2 7.533

The number of transformers for each rated power is close to the probabilities set by the
reference indicators in Figure 4-12.

Medium to low voltage transformers

Voltage (kV) Rated Power (kVA) Number
20/0.4 1000 34
20/0.4 630 52
20/0.4 400 40

41
The types of MV electrical lines are MV_U_1/2, which are underground.

Medium voltage network

Voltage (kV) Type ID Length (km)

20 MV_U_1 27.163
20 MV_U_2 4.032

Figure 4-7 shows the five (colored) MV feeders. The HV/MV substation is represented by
a triangle. Each feeder is painted with a different color. Black lines represent loops in the
MV network normally open for increasing the reliability of supply in case of network
outages.

Figure 4-7 Urban MV feeders

The following table presents information broken down per MV feeder.

42
 Feeders parameters

LV network [km] MV network [km] MV/LV Substations

MV Feeder #1 8.00 4.97 15
MV Feeder #2 11.49 5.86 25
MV Feeder #3 11.81 6.21 26
MV Feeder #4 11.01 5.34 31
MV Feeder #5 11.57 6.98 29
Loops 0.00 1.84 0

High to medium voltage substations

Rated Power
Voltage (kV) (MVA) No.

132/20 80 1

4.2.1.2 Semi-urban (#2)

Figure 4-8 shows the large-scale network with a semi-urban configuration, including MV
feeders (blue lines), LV feeders (black thin lines), as well as MV/LV substations (red
circles) and the HV/MV substation (blue triangle). It represents the outskirts of a city.

Figure 4-8 Semi-urban network

The proportion of MV to LV consumers is similar to the urban area case. In the semi-
urban area there are however some MV overhead electrical lines. Again, the total
installed capacity in MV/LV substations and HV/MV substations are close.

43
 Aggregated inputs

No. Peak Power (MW)

LV Consumers 13998 68.5
MV Consumers 40 6

Overhead Underground
LV km 67.21 47.75
MV km 13.24 37.16

No. Capacity (MVA)

MV/LV subs. 161 97.91
HV/MV subs 1 120

All the types of electrical lines defined in section 4.2.2.6 are used in the semi-urban
area, including overhead-pole, overhead-façade and underground.

Low voltage network

Voltage (kV) Type ID Length (km)

0.4 LV_IO_1 7.43
0.4 LV_IO_2 1.217
0.4 LV_UO_1 32.25
0.4 LV_UO_2 26.314
0.4 LV_UU_1 43.128
0.4 LV_UU_2 4.623

The rated power of MV/LV transformers in the semi-urban area covers a wider spectrum
than in the urban area, in which sizes where higher.

Medium to low voltage transformers

Voltage (kV) Rated Power (kVA) No.

20/0.4 1000 38
20/0.4 630 57
20/0.4 400 54
20/0.4 250 8
20/0.4 100 4

Types MV_O_1/2 refer to overhead, while types MV_U_1/2 refer to underground.

Medium voltage network

Voltage (kV) Type ID Length (km)
20 MV_O_1 12.033
20 MV_O_2 1.204
20 MV_U_1 36.296
20 MV_U_2 0.859

44
Figure 4-9 shows the ten MV feeders. The HV/MV substation is represented by a triangle.
Each feeder is painted with a different color. Black lines represent loops in the MV
network.

Figure 4-9 Semi-urban MV feeders

The following table shows information broken down per MV feeder.

LV network MV network MV/LV

[km] [km] Substations

MV Feeder #1 11.58 4.35 16

MV Feeder #2 12.39 4.94 16
MV Feeder #3 10.17 3.43 16
MV Feeder #4 11.81 4.49 16
MV Feeder #5 9.33 3.31 15
MV Feeder #6 9.51 2.50 10
MV Feeder #7 9.59 3.95 12
MV Feeder #8 19.94 8.59 30
MV Feeder #9 14.19 8.18 21
MV Feeder #10 6.45 1.97 9
Loops 0.00 4.67 0

High to medium voltage substations

Rated Power
Voltage (kV) (MVA) Number
132/20 120 1

45
4.2.1.3 Rural (#3)
Figure 4-10 shows the rural large-scale network, including MV feeders (blue lines), LV
feeders (black thin lines), as well as MV/LV substations (red circles) and a HV/MV
substation (blue triangle). It represents a distribution network supplying electricity to
several small settlements and farms in the countryside.

Figure 4-10 Rural network

As opposed to the previous distribution networks, in the rural network most of the
electrical lines are overhead, both in MV and LV.

Aggregated inputs

No. Peak Power (MW)

LV Consumers 7727 25.28
MV Consumers 20 3

Overhead Underground
LV km 201.97 9.49
MV km 111.24 19.88

No. Capacity (MVA)

MV/LV subs. 152 40.25
HV/MV subs 1 80

46
As in the semi-urban network, all the types of LV electrical lines are used in the rural
network, being LV_IO_1/2 (overhead-pole) the most common one.

Low voltage network

Voltage (kV) Type ID Length (km)

0.4 LV_IO_1 96.994
0.4 LV_IO_2 25.732
0.4 LV_UO_1 66.871
0.4 LV_UO_2 12.476
0.4 LV_UU_1 9.384
0.4 LV_UU_2 0.124

Only small size transformers (up to 400kVA) are used in the rural area.

Medium to low voltage transformers

Voltage (kV) Rated Power (kVA) Number

20/0.4 400 42
20/0.4 250 79
20/0.4 100 29

Most MV electrical lines are overhead, being type MV_O_1 the most common one.

Medium voltage network

Voltage (kV) Type ID Length (km)

20 MV_O_1 107.608

20 MV_O_2 3.624

20 MV_U_1 19.885

Figure 4-11 shows the six MV feeders. The HV/MV substation is represented by a
triangle. Each feeder is painted with a different color. Black lines represent loops,
normally open, for increasing reliability of supply in the MV network.

47
 Figure 4-11 Rural MV feeders

The following table shows information broken down per MV feeder.

LV Network MV Network MV/LV

[km] [km] Substations
MV Feeder #1 11.54 9.95 6
MV Feeder #2 21.61 15.21 10
MV Feeder #3 43.94 24.93 33
MV Feeder #4 26.26 10.00 25
MV Feeder #5 67.51 38.09 47
MV Feeder #6 40.73 14.95 31
Loops 0.00 17.99 0

High to medium voltage substations

Rated Power
Voltage (kV) (MVA) Number
132/20 80 1

4.2.2 Feeder-type network topologies

These topologies include two sub-categories: MV feeders and LV feeders.
The MV feeder networks (#4-11) have been built to be able to analyze the impact on the
continuity of supply (i.e. duration and frequency of supply interruptions) depending on
the network configuration and on the degree of automation of the distribution network.
Therefore, for each network, two versions are being provided, one network with a low
degree of automation and another one with a high degree of automation. Under a low
degree of automation all the protection equipment (switches, breakers, etc.) are
manually operated. Under a high degree of automation, some of the switches and
breakers are remotely controlled. Typically, the tele-controlled protection equipment is
regularly spaced along the MV feeders to maximize reliability improvements.

48
Also the MV urban and semi-urban topologies have been built taking into account the
European DSO indicators. The main parameters required for this task have been: the
feeder lengths, the number of MV/LV substations per HV/MV substation and the MV/LV
transformer substation capacity. These indicators were broken down in urban and rural.
Unfortunately only a fraction of DSOs provided this information, so that these models
have been built using the available information.

The MV/LV transformer capacity is used to determine the transformation capacity of

each MV/LV transformer substation. As it is shown in Figure 4-12 and Figure 4-13, the
typical transformation capacity8 is higher in urban networks than in rural network.

Figure 4-12 Typical transformation capacity of the MV/LV secondary substations in urban areas

Figure 4-13 Typical transformation capacity of the MV/LV secondary substations in rural areas

8
To improve the readability of the document we decided to report here the two figures already presented in
section 3.2.1.

49
The average number of MV/LV substations per feeder is used to determine how many
MV/LV substations are connected at each feeder. As it is shown in Figure 4-14 and
Figure 4-15, the number of MV/LV substations per feeder is higher in rural areas than in
urban areas.

Figure 4-14 Average number of MV/LV substations per feeder in urban areas

Figure 4-15 Average number of MV/LV substations per feeder in rural areas

50
The MV feeder length is used to determine the total length of the feeders in the
representative network, which indirectly results in a distance between MV/LV
substations. As it is shown in Figure 4-16 and Figure 4-17, the typical MV feeder length
is higher in rural areas than in urban areas.

Figure 4-16 Average length per MV feeder in urban areas

Figure 4-17 Average length per MV feeder in rural areas

51
4.2.2.1 Urban MV network: two substations interconnected (#4 &
#5)
This network represents two HV/MV substations with feeder support connecting each
other. This configuration is representing MV feeders in an urban MV network.

Figure 4-18 Two substation urban network

The representative MV network comprises two substations connected to buses 1 and 25,
respectively. Two feeders per substation were modeled, with 10, 11, 12 and 13 MV/LV
substations each. Two normally open loop branches (24-48, 14-37) connect the end of
the feeders, respectively. Therefore, in case of failure of any feeder branch, one
substation can supply part of the loads of the other substation. This configuration would
also cope with HV/MV transformer failures as every feeder can be supplied through two
alternative HV/MV substation transformers.

The feeder length and the number of MV/LV substations, as well as the distribution
function of the transformer capacity has been selected based on the European DSO
indicators.

Feeder length MV/LV

[km] Substations
MV Feeder #1 4.2 13
MV Feeder #2 4.0 10
MV Feeder #3 4.5 12
MV Feeder #4 4.5 11
Loop branches 0.6 0

52
 Figure 4-19 MV/LV Transformer capacity

In the buses modeling the MV/LV transformers, it has been assumed a load of 75% of
the capacity of the transformer, with a 0.95 power factor. A power flow has been run
with these parameters, and the operational constraints have been verified.

53
4.2.2.2 Urban MV network: one substation and one switching station
(#6 & #7)
This network represents three MV feeders of a HV/MV substation. The ends of the MV
feeders are connected to a switching station. This configuration is also representative of
urban MV networks.

The HV/MV substation is connected to bus 1. There are three feeders with 10, 11 and 12
MV/LV substations each. The ends of the feeders are connected to a switching station in
bus 35, through normally open branches (34-35, 23-35, 13-35). There is a support
feeder connected to the switching station (1-35), to provide an alternative way to supply
electricity in the case of any feeder branch failure. This support feeder has been
represented as normally closed (instead of normally open), so that bus 35 is not isolated
from the rest of the other buses and the power flow can be executed 9.

Figure 4-20 Urban switching station

In this case, the reference indicators are 3.9 km feeder length, 11 MV/LV substations per
feeder, and 400kVA, 630kVA and 1000kVA as the MV/LV transformation capacity, being
630kVA the most common one.

Feeder length MV/LV

[km] Substations
MV Feeder #1 4.01 12
MV Feeder #2 3.69 10
MV Feeder #3 4.21 11
Support feeder 0.42 0
Loop branches 0.83 0

9
Otherwise bus 35 would be isolated and its voltage would not be defined.

54
 Figure 4-21 MV/LV Transformer capacity

In the buses modeling the MV/LV transformers, it has been assumed a load of 75% of
the capacity of the transformer, with a 0.95 power factor. A power flow has been run
with these parameters, and the convergence of the power flow has been tested,
checking voltage and thermal limits.

4.2.2.3 Semi-urban MV network: substation ring (#8 & #9)

This network represents a substation ring with two feeders of a HV/MV substation
connected by a loop. It is modeling a semi-urban MV network.

Figure 4-22 Semi-urban substation ring

55
The transformer of the HV/MV substation is connected to bus 1. There are two main
feeders with 16 and 18 MV/LV substations each. A few secondary feeders are radially
connected to the main feeders. The two ends of the feeders are connected by a normally
open loop branch (19-39), which allows recovering part of the load in the case of any
feeder branch failure.
The feeder length and the number of MV/LV substations per feeder have been increased
compared to the urban networks, to take into account that according to the European
DSO indicators, both indicators increase in rural networks. The values are an
intermediate point between the values of the urban and rural indicators. The MV/LV
transformer capacity is also reduced compared to the ones in the urban networks, but
not as much as in rural networks.

Feeder MV/LV
length [km] Substations
MV Feeder #1 6.08 16
MV Feeder #2 7.87 18
Loop branch 0.10 0

Figure 4-23 MV/LV Transformer capacity

In the buses modeling the MV/LV transformers, it has been assumed a load of 75% of
the capacity of the transformer, with a 0.95 power factor. A power flow has been run
with these parameters, and the convergence of the power flow has been tested,
checking voltage and thermal limits.

56
4.2.2.4 Rural medium voltage network (#10 & #11)
This representative rural MV network represents a HV/MV substation with several radial
feeders, some of them connected by loop branches (black lines).
This network has been built using the Reference Network Model, first modeling both the
LV & MV consumers and feeders, and then extracting only the MV network. Nearby
consumption points are modeling small settlements and isolated consumptions are
representing farms.
The HV/MV substation is represented by a triangle. Each feeder is painted with a
different color. The black lines represent loop branches normally open connecting
different MV feeders.

Figure 4-24 Rural MV network

The following tables and figures show the network indicators.

Network Indicator Value

MV circuit length per MV Supply Point 0.705
MV underground ratio 0
Number of MV Supply Points per HV/MV substation 115

57
 Feeder
length MV/LV
[km] Substations
MV Feeder #1 12.00 17
MV Feeder #2 9.31 15
MV Feeder #3 6.11 3
MV Feeder #4 6.84 15
MV Feeder #5 6.38 17
MV Feeder #6 4.07 1
MV Feeder #7 20.42 42
MV Feeder #8 13.48 5
Loop branches 2.42 0

Figure 4-25 MV/LV Transformer capacity

The network comprises 118 buses. Therefore, only aggregated data of the rural medium
voltage network is presented in this section.

Two protection equipment tables corresponding to a low and high degree of automation
are available on the SESI webpage10. When the degree of automation is low (#10), all
protection equipment are manually operated. When the degree of automation is high
(#11), there are tele-controlled switches in buses 31, 42 and 55; as well as tele-
controlled breakers at the beginning of the feeders.

10
http://ses.jrc.ec.europa.eu/distribution-system-operators-observatory

58
 Aggregated inputs

No. Peak Power (MW)

MV Supply Points 115 19.9

Overhead Underground
MV km 81.03 0

No. Capacity (MVA)

HV/MV subs 1 80

Medium voltage network

Voltage
(kV) Type ID Length (km)

20 MV_O_1 81.03

High to medium voltage substations

Rated Power
Voltage (kV) (MVA) No.

132/20 80 1

4.2.2.5 LV feeders
The representative LV feeder networks were built with the aim of analyzing the impacts
of DG on single LV networks, such as the installation of photovoltaic distributed
generation units. Network #12 models an urban LV grid while network #13 models a
semi-urban LV grid. In the semi-urban area, distances are longer and there are more LV
consumers per feeder.

4.2.2.5.1 Urban low voltage network (#12)

This network represents a MV/LV substation with LV feeders. This configuration is
modeling a high density urban LV network. This network has been built using the
Reference Network Model.

Figure 4-26 Urban LV network

59
The following table shows the network indicators.

Network indicator Value

LV network length per LV Consumer 0.0024
LV underground ratio 1
Number of LV consumers per MV/LV substation 107
LV consumer peak power [kVA] 5.9

4.2.2.5.2 Semi-urban low voltage network (#13)

This network represents a MV/LV substation with LV feeders. This configuration is
modeling a semi-urban LV network. It is based on representative feeder 2 in (Rigoni
2014).

Figure 4-27 Semi-urban LV network

The following table shows the network indicators.

Network indicator Value

LV circuit length per LV consumer 0.011

LV underground ratio 1
No. of LV consumers per MV/LV substation 108
LV consumer peak power [kVA] 3.70

This network is composed of 115 buses. The following tables show the aggregated data.

60
 Aggregated inputs

No. Peak Power (MW)

LV Consumers 108 0.259

Overhead Underground
LV km 0 1.154

No. Capacity (MVA)

MV/LV subs. 1 0.400

Low voltage network

Length
Voltage (kV) Type ID (km)
0.4 LV_UU_2 1.154

Medium to low voltage substations

Voltage (kV) Rated Power (kVA) Number

20/0.4 400 1

4.2.2.6 Standard equipment used to build the networks

A brief summary of the standard equipment used to build the representative networks is
given in the following.

In particular, Table 4-4 shows the parameters of the LV feeders, Table 4-5 shows the
characteristics of the MV/LV transformers, and Table 4-6 shows the parameters of the
MV feeders. R and X are the resistance and reactance per phase of the power lines
respectively in ohms/km. B is the susceptance in ohms-1/km. Rsc and Xsc are the short-
circuit resistance and reactance of the transformers in unitary units referred to the rated
power and voltage of the respective transformer.

Rated
Voltage current R X
(kV) Type ID Name Type (A) (ohms/km) (ohms/km)

0.4 LV_IO_1 3x 50mm2 Overhead - pole 150 0.65 0.1

2
0.4 LV_IO_2 3x 150mm Overhead - pole 305 0.2 0.1
2
0.4 LV_UO_1 3x 25mm Overhead - façade 100 1.4 0.1

0.4 LV_UO_2 3x 95mm2 Overhead – façade 230 0.32 0.1

0.4 LV_UU_1 3x 95mm2 Underground 255 0.39 0.075

2
0.4 LV_UU_2 3x 240mm Underground 420 0.14 0.08
Table 4-4 Low voltage feeders

61
 Voltage Rated Power Rsc Xsc
(kV) (kVA) (p.u.) (p.u.)

20/0.4 1000 0.012 0.04

20/0.4 630 0.012 0.04
20/0.4 400 0.012 0.04
20/0.4 250 0.012 0.04
20/0.4 100 0.012 0.04

Table 4-5 Medium to low voltage transformers

Rated
Voltage Type current R (ohms X (ohms B (ohms1
Name Type
(kV) ID /km) /km) /km)
(A)
MV_O LA80
20 Overhead 250 0.42 0.39
_1 (60-AL1/14- -
STIA)
LA180
MV_O
20 (147- Overhead 425 0.18 0.36
_2 -
AL1/34-
ST1A)
MV_U Undergrou
20 3x 120mm2 300 0.21 0.11 7.57E-05
_1 nd
MV_U Undergrou
20 3x 400mm2 515 0.07 0.09 1.17E-04
_2 nd

Table 4-6 Medium voltage feeders

Table 4-7 shows the parameters of the high to medium voltage substations.

Rsc Xsc
Rated Power
Voltage (kV) (MVA) (p.u.) (p.u.)

132/20 80 0.003 0.10

Table 4-7 High to medium voltage substations

The parameters of the equipment are part of the RNM catalogue (R. Cossent 2011),
which is based on manufacturer data11 12 13.

The reactance of the MV overhead power lines has been obtained with the formula (1).
X=2πf(0.5+4.605 Log D/r)x10-4 ohms/km (1)

where: X: Reactance in ohms/km

11
Nexans, 6-36Kv Medium Voltage Underground Power Cables.
12
Ormazabal, Distribution Transformers up to 5MVA.
13
Eaton, Power Distribution Systems

62
f: Frequency in Hertz = 50
D: Geometric average distance between line conductors in mm.
r: Conductor radius in mm.

63
64
5 Evaluation of policy options based on RNM

This section presents some simulation examples using the built representative networks
to analyse the impact of DER penetration and network automation on the technical
performance of these distribution networks. These analyses have been mainly done to
present illustrative examples of the potential applications of the representative networks
built in the DSOs Observatory project. 24 hour profiles were modelled or assumed for
consumers and for the distributed generation connected to the networks. Power flows for
each hour were computed as well.
In the analysis, the impact of two RES technologies was analyzed: solar and wind. In
particular, the penetration levels expected in years 2020 and 2030 were considered.
Figure 5-1 shows the expected solar penetration14 in the EU Member States, according to
the study prepared by the European Photovoltaic Industry Association (EPIA) (EPIA
2012). It is based on new EPIA scenarios for the penetration of PV electricity in 2030:

- the Baseline scenario envisages a business-as-usual case with 4% of EU

electricity demand provided by PV in 2020 and 10% in 2030;

- the Accelerated scenario foresees PV meeting 8% of the demand in 2020 and

15% of the demand in 2030 and based on current market trends;

- the Paradigm Shift scenario is based on the assumption that all barriers are lifted
and that specific boundary conditions are met, which foresees PV supplying up to
12% of EU electricity demand by 2020 and 25% in 2030.

Figure 5-1 Expected solar PV penetration in 2020 and 2030 (EPIA)

14
It is measured as the percentage of EU electricity energy demand provided by PV.

65
In our analysis, the penetration of solar generation is set according to the accelerated
scenario. This scenario was selected because it is believed to represent a realistic view of
market development in Europe until 2030 given the targets fixed in the 2030 framework.
It roughly assumes that the same market conditions observed in 2011 will be maintained
throughout the coming two decades. The amount of electricity that a PV installation
generates clearly depends on the weather conditions. To this aim three types of days can
be considered: rainy, cloudy and sunny. In order to model the most stressing situation
for the network, the condition of peak generation, i.e. the sunny day, was selected.

In the case of wind generation, it is important to take into account that only a fraction of
the energy is produced onshore, and not all of that is connected to the distribution
networks. According to scenarios described in the literature (J. Wilkes 2014) (EWEA
2011), the expected wind penetration was set to 14.9% in 2020 and 28.5% in 2030,
80% of that onshore, and 57.4% connected to MV (which corresponds to 6.8% and
13.1% for 2020 and 2030 respectively). Regarding hourly profiles, there is much
variability in the energy that can be produced by a wind turbine. The wind production of
a particular installation15 was selected in order to analyze a real wind hourly production
(EnerNex 2011). Three types of profiles were identified corresponding to a minimum,
middle and maximum production days. The maximum production day was selected in
order to consider the condition of peak generation as the more stressing situation for the
distribution network. Both the large-scale rural and urban representative networks were
selected to carry out the impact analyses. In rural networks, due to the existence of
longer distances, voltage problems are the most stringent operational constraints, while
in urban networks, with shorter distances and higher density of consumers, congestions
are the most frequent operational problem.

5.1 Selected scenarios

The main scenarios considered were a demand only scenario with no RES (Sc. 0 in the
following), and two more scenarios, corresponding to the expected RES penetration16 in
years 2020 (Sc. 1 in the following) and 2030 (Sc. 2 in the following). As shown in Figure
5-2 an additional scenario (Sc. 3 in the following) has been considered, corresponding to
twice the DG penetration in 2030 (2030 x2). The penetration levels in 2020 and 2030
are based on the literature, as previously described.

The PV panels were modeled as located on residential consumers, and each PV unit was
placed in the same location of an existing consumer. When setting the annual energy
production of a PV unit, an onsite PV size limit was considered. This onsite PV size limit
was interpreted as a restriction on the size of the PV, based on the annual consumption
of the consumer. In particular, in the rural network, it was modeled that the annual
energy production of the PV units could be 50%, 100% and 200% of the respective
consumed energy. For example 50% onsite PV size limit means that each PV unit is
annually producing 50% of the consumer annual energy (only in the consumer premises

15
Wind data was obtained from the National Renewable Energy Laboratory (NREL), which is operated by the
Alliance for Sustainable Energy, LLC (ALLIANCE) for the U.S. Department of Energy (DOE).
16
The penetration is measured as the percentage of EU electricity energy demand provided by PV and wind,
respectively.

66
in which they are installed). On the other hand, in the urban network, the onsite PV size
limit was extended to a finer range comprised between 60% and 200%.

Figure 5-2 RES penetration scenarios

5.2 Results of the simulations

A power flow was run for each hour of the day. The 24 power flows were classified into
three time tiers, 1) 10h-14h corresponding to peak PV generation hours, 2) 18h-20h
corresponding to peak demand hours, and 3) the rest of the hours.

Residential consumers have much variability in their consumption, having usually short
periods of peak demand and a lower demand during the rest of the time. Therefore, in
the case of these customers, there is a significant difference in using the average
consumption versus using real individual consumptions, and this can have a considerable
impact on the power flow results. In order to capture the variability of LV residential
consumers and to get more realistic results, the Load Profile Generator was used to
generate the hourly profiles of residential consumers (Pflugradt 2013). A total of 72
profiles were obtained and they were assigned to individual consumers. Each of these
profiles contains 24 values, one for each hour. A separated power flow was then run for
each hour.

Commercial load profiles are based on (Jenkins 2014). Four commercial load profiles
were considered in the analysis. In the case of the commercial consumers there is
usually a stable consumption during working hours. Moreover, the total number of
commercial consumers connected to the representative distribution networks is much
smaller than the total number of residential consumers. Therefore, in this case,
capturing the heterogeneity in the consumption is less important, and fewer profiles
have been considered in the analysis.

67
5.2.1 Impact of solar PV on distribution networks
In each power flow, the reference voltage at the slack bus was set with the criteria of
fixing the median of the bus voltages in the network to 1.0 p.u. Therefore, half of the
buses had a voltage lower than 1.0 p.u. and half of the buses had a voltage higher than
1.0 p.u. In this way, voltages were controlled as close as possible to 1.0 p.u. in order to
minimize voltage impacts.

Figure 5-3 shows the bus voltages in the network during the peak PV hours for the
100% onsite PV size limit (that is, each installed PV is producing all the energy that the
consumer requires during the year) in the four scenarios. As shown in Figure 5-3, most
of the bus voltages are in the range 0.95 - 1.05 in the first three scenarios (Sc. 0. Sc. 1
and Sc. 2) but present a significant higher voltage spread than Sc. 0, being the voltage
spread slightly higher in Sc. 2 than in Sc. 1.
Finally, it can be observed that there are many more buses suffering under-voltages and
over-voltages conditions in Sc. 3 when compared with the other scenarios. Despite the
fact that the allowed operational voltage limits are ±10% of the nominal values
according to standard EN 50160, the planning criteria followed by DSOs are usually more
restrictive. For example, in (Long Island Power Authority 2010) voltage drops are limited
to ±5%. These security margins are necessary because, due to the traditional passive
operation paradigm, the voltage spread in operating conditions can increase significantly
with respect to the considered planned conditions.

25
Sc. 0
Sc. 1
20 Sc. 2
Sc. 3
Frequency (%)

15

10

0
0.9 0.95 1 1.05 1.1
Voltage (p.u.)
Figure 5-3 Bus voltages in the rural network for a 100% PV size limit (Peak PV hours)

68
Figure 5-4 shows the bus voltages during the peak demand hours. During these hours
(18-20 h) the PV generation is almost negligible and then the bus voltages are very
similar in all the scenarios, not being significantly influenced by the PV production.

25
Sc. 0
Sc. 1
20 Sc. 2
Sc. 3
Frequency (%)

15

10

0
0.9 0.95 1 1.05 1.1
Voltage (p.u.) .
Figure 5-4 Bus voltages in the rural network for a 100% PV size limit (Peak demand hours)

Figure 5-5 shows the voltage spread in the rural network depending on the scenario and
on the onsite PV size limit implemented. The voltage spread is measured as the
difference between the maximum voltage and the minimum voltage across the network
in per unit (p.u.). Only with a 50% onsite PV size limit voltage spreads are below the
±10% limit (0.2p.u.) in Sc. 1 and in Sc. 2. As the DG penetration increases, the voltage
spread increases. The voltage spread depends strongly on the onsite PV size limit.

Figure 5-5 Voltage spread in the rural network

69
5.2.1.1 Economic impact assessment of RES penetration on network voltages
In order to monetize the impact of RES penetration on network voltages, we have used a
penalty cost function. This penalty function is used mainly to quantify from an economic
perspective the impact of voltage spreads on the considered network, as shown in Figure
5-6. It assumes a cost of zero when the bus voltage is within the range 0.95 - 1.05 p.u.;
a 3€/kWh cost 17 for voltages outside de ±10% limit; and a quadratic function for the
intervals 0.9 - 0.95 p.u. and 1.05 - 1.1 p.u. (a graphical representation is shown in
Annex D: Cost functions). From the graph two main causes can be identified as
responsible for increasing network costs due to voltage levels outside operational limits.
One of these is the penetration of RES, which increases voltage costs as it can be
observed by comparing the results obtained for the different scenarios. The size of the
PV units would be a second cause of these costs, which could be limited by devising a
more efficient onsite PV size limit strategy. This effect can be observed when comparing
the different onsite PV size choices, within a given scenario. In some cases, the impact
of the onsite PV size limit on voltage costs is even greater than the impact due to an
increase in RES penetration. For instance, Sc. 1 with a 200% onsite PV size limit has a
higher voltage cost than Sc. 2 with a 100% onsite PV size limit. This means that, in this
particular case, the question is not only how much renewable generation is connected to
the system, but also how this is distributed throughout the whole network.

Figure 5-6 Voltage cost impact on the rural network

Figure 5-7 and Figure 5-8 shows through the colored dots where the voltage levels are
outside the established limits in the rural network. Two snapshots corresponding to
14:00 p.m. in Sc. 0 and in Sc.1 have been selected. Nodes with voltages below 0.95 p.u.
(undervoltages) are identifed by red circles, and with voltages above 1.05 p.u.
(overvoltages) by blue circles. Figure 5-7 shows that in Sc. 0 only some nodes with
undervoltages can be identified.

17
This cost is associated to the cost of the non-served energy due to a consumer supply interruption.

70
 Figure 5-7 Voltage map of the rural network at 14:00 in Sc. 0

Figure 5-8 shows the voltage map in Sc. 2 with a 200% onsite PV size limit. In some
buses in which there were low voltages in Sc. 0, such problems have spread to the
nearby buses. In other cases, the local generation have raised voltages, eliminating the
previous undervoltage problems.

Both undervoltages and overvoltages occur inside settlements and outside settlements.
Even though PV generation was located randomly on the network, the undervoltages and
overvoltages do not tipically appear in the same area. Instead, they are normally
concentrated in some areas. For example, in this rural network there are two
settlements with undervoltages, one settlement with overvoltages, and one settlement
which simultaneously has undervoltages and overvoltages. In those cases some planning
remedial actions can be proposed to improve the network performances. For instance, in
the settlements which only have one type of problem, undervoltages or overvoltages,
voltage regulators or transformers with on-load tap changers could be installed to
alleviate those voltage problems.

Figure 5-8 Voltage map of the rural network at 14:00 in Sc. 2, with a 200% onsite PV size limit

71
Figure 5-9 shows the aggregated load/generation profiles of consumers, PV, and wind in
Sc. 2. The values are positive for consumption and negative for generation. The total net
demand is near zero for the peak PV hours. In Sc. 3, with even higher DG penetration
levels, the net demand becomes negative in peak PV hours, which implies the existence
of reverse power flows in the HV/MV substation during those hours. Such conditions
need to be avoided since they can cause high losses in the system and can have an
impact on protection equipment.

Figure 5-9 Aggregated profiles in year 2030

The aggregated profiles in Figure 5-9 are the same for each scenario independently of
the size of the PV units (i.e. independently of the onsite PV size limit). This means that
with a 50% onsite PV size limit, more PV units would have to be installed to reach the
same total penetration level than with a 100% onsite PV size limit.

The percentages of consumers which have installed a PV in each scenario and onsite PV
size limit are shown in Table 5-1. For instance, the 50% onsite PV size limit, for a global
level of PV penetration of 8% (Sc. 1), means that 17% of total consumers install PV
producing the 50% of their total electricity energy demand. The percentage of
consumers in Sc. 1 and Sc. 2 installing PV for a 100% onsite PV size limit is 8.8% and
16%.

SC. 0 SC. 1 SC. 2 SC. 3

50% onsite PV size limit 0.0% 17.0% 31.3% 57.6%

100% onsite PV size limit 0.0% 8.8% 16.1% 31.3%

200% onsite PV size limit 0.0% 4.4% 8.3% 16.1%

Table 5-1 Percentage of consumers installing PV for each scenario and onsite PV size limit

72
Thus, the expected penetration levels of 2020 and 2030 can be achieved with different
onsite PV size limit choices. When deciding to implement a given onsite PV size limit
choice one has to take into account that different choices can have different impact on
the network and on the needed investments. Smaller PV units have less negative impact
on the network performances but more consumers need to be encouraged to install PV
to reach the same penetration level. Additionally, a higher number of active network
users could require a bigger effort for DSOs in terms of management and operation
efficiency. This has implications, for instance, regarding the obligations set on generators
regarding observability and/or controllability which are frequently defined depending on
the size of DG units. For example, if one sets an observability requirement for all units
above 1MW, and assuming a total installed capacity of 10MW, it becomes relevant
whether such capacity corresponds to 10 units of 1MW (thus observable) or 100 units of
0.1MW (which are not observable according to the hypothesized requirements). On the
other side, big size (and less sparse) units create more problems to the network but a
smaller amount of them is needed to reach the established penetration targets. As
conclusion of this analysis, it can be observed that the assessment of RES penetration on
network impacts cannot be derived only from the aggregated load/generation profiles,
such as the one shown in Figure 5-9. As emphasized, the number of PV units and their
concentration/distribution on the network, given by the percentage of consumers with PV
units (in our case), is also a relevant parameter, which has been modeled through
several onsite PV size policies. As already shown in Figure 5-6, the expected network
impacts (in terms of costs) would highly depend on those onsite PV policies. Figure 5-10
shows the number of overloads 18 in the rural network in each scenario and for each
onsite PV size limit. As it happened with bus voltages, the RES penetration and the
onsite PV size limit are causes for congestions. In this particular case, there are no
overload problems in any scenario with a 50% onsite PV size limit. Some overloads occur
for a 100% onsite PV size limit, and the number of overloads increases significantly for a
200% onsite PV size limit in the three scenarios. Despite seeming unrealistic, the limit of
200% onsite PV size could be driven by generous net-metering policy with additional
compensation for excess production or by community net-metering programs (which
allow compensating the consumption of some customers with the excess production of
their neighbours/associates).

18
The number of overloads is calculated as the network branches in which the power flow exceeds the
maximum allowed operational limit times the number of hours in which this condition happens.

73
 Figure 5-10 Number of overloads in the rural network

5.2.1.2 Economic impact assessment of RES penetration on network overloads

As previously done, the impact of RES penetration on network overloads have been
monetized by means of a penalty cost function. The utilization factor can be defined as
the power through the branch (conductor or transformer) divided by the branch
capacity19. When the utilization factor of a network branch is below one the branch is not
overloaded and the associated penalty cost is zero. For higher utilization factors the
penalty cost increases linearly with a slope of 3€/kWh, which is related to the cost of the
non-served energy in case of an interruption. Figure 5-11 shows the calculated network
overload costs. Again, the onsite PV size limit is critical. Significant overloads only
appear in the rural network for a 200% onsite PV size limit, meaning that overloads in
the rural network only occur for high PV sizes.

Figure 5-11 Network overload costs in the rural network

19
Despite its name, some power lines require a higher capacity to deal with voltage problems, and therefore in
that case a low utilization factor doesn’t mean that the line is infra-utilized.

74
In order to analyse overloads, it was deemed relevant to perform a similar analysis on
the urban network. Since this network shows a higher concentration of consumers, it is
less prone to voltage problems than it is to overloads. The scenario analysed
corresponds to Sc. 2, with an onsite PV size limit in the range of 60-200%. Figure 5-12
shows the number of overloads in the urban network depending on the onsite PV size
limit. For a 60% limit there are no overloads. Starting at 80% onsite PV size limit, some
overloads appear in the network. As the onsite PV size limit increases, that is per each
consumer with installed PV a greater percentage of energy is produced, overloads
increase exponentially.

Figure 5-12 Number of overloads (branch*hour)

By comparing this result with the ones in the previous section it can be concluded that
voltage problems occur in the rural network even for very low onsite PV size imposed
limits. On the contrary, overloads only appear in urban networks when the size of PV
installations is much larger.

5.2.2 Storage units to mitigate voltage spreads in the network

Figure 5-13 shows the voltage spread in the rural network in Sc. 1 for a 100% onsite PV
size limit, when storage is installed to mitigate the effect of distributed generation. The
starting point, with no storage installed, is over the ±10% limit (0.2 p.u.). When storage
is installed the voltage spread decreases. 28kWh storage units would be the limit to keep
the voltage spread below 0.2p.u. Storage units were connected to buildings with a
previous PV installation, instead of being designed for individual consumers. As it can be
observed in Figure 5-13, even for 100kWh storage units, the voltage spread would be
only slightly below the ±10% limit.

75
 0.5

0.4

Voltage spread (p.u.)

0.3

0.2

0.1

0
0 20 40 60 80 100
Storage capacity (kWh)

Figure 5-13 Voltage spread in Sc. 1 with increased storage unit capacity

This analysis highlights the advantages of increasing the allowed voltage spread,
whenever it is technically feasible (i.e. if it does not harm or result in a wrong operation
of the installed equipment and consumer devices). Especially in countries in which the
voltage spread is currently more restrictive than the ±10% of EN 50160, the relaxation
of this constraint could significantly facilitate the integration of more distributed
generation.

Figure 5-14 shows how the network voltage and the overload costs decrease as storage
capacity increases in storage units. In this graph the overload cost is low, as there are
little congestions in this scenario (this can be checked in Figure 5-10 and Figure 5-11).
Besides, it can be observed that the marginal benefit of storage decrease with the
storage size.
Storage was modeled as only installed in PV premises, not in wind park sites. Therefore,
even for 100kWh storage, there is a significant remaining voltage cost (in Sc. 0 the
voltage cost was 370 €/day).

Figure 5-14 Voltage and overload cost in Sc. 1, for a 100% onsite PV size limit, with increased
storage unit capacity

76
The high costs per kWh20 for batteries suggest however that other solutions should be
considered to mitigate voltage spread in the distribution network.

5.2.3 Reliability analysis

In the following, a reliability analysis is carried out (Rodriguez, et al. 2016) to calculate
the reliability indexes in the selected representative networks. The model simulates a
fault that leads to an outage for each of the electrical power lines of the network, and
then models the operation of the switches and breakers to locate and isolate the fault
and to restore the service.

The sequence of steps carried out for each fault consists of:

1. Determining affected loads

2. Applying smart grid solution
3. Sending maintenance crews
4. Manual switching
5. Visual inspection
6. Reparation
Through this approach, the customers’ interruptions and the relative durations are
calculated. By aggregating these numbers for all the possible faults, SAIDI21 and SAIFI22
indexes are gathered.

20
At the moment there are different kind of technologies for storage and several estimates on their prices.
Generally those solutions that are considered scalable are still far for being considered cheap (Luo 2015).
21
SAIDI is the System Average Interruption Duration Index. It measures the duration of the interruptions.
22
SAIFI is the System Average Interruption Frequency Index. It measures the frequency of the interruptions.

77
 In brief, the time to restore service in a feeder is comprised for each load from some of all
of the following elements:
t = toperator + nsteps,j * toperator,j + t1 + Σsteps,i [ t2,i+t3,i ] + t4 + t5 ,

where:
toperator is a fixed term to account for the response time of the operator in the control centre
(min)
nsteps,j is the number of switching actions performed by the operator in the control centre
to isolate the faulty segment among two telecontrolled elements
toperator,j is a fixed term to account for the time required for each switching action performed
by the operator in the control centre (min)
t1 is a fixed term to account for the time required for fault detection and sending a
maintenance crew (min)
t2 is a fixed time for each step of the dichotomic search process (operation of the
switches, getting into the car, etc) (min)
t3,i is a variable time proportional to distance to travel in step i, considering a certain
speed s3 (min)
t4 is a variable time for fault localization along a segment, proportional to distance to
cover, considering a certain speed s4 (min)
t5 is a fixed time to repair a fault in a branch of the feeder (min)

Figure 5-15 Parameters used in the simulations to obtain SAIDI and SAIFI

Three network configurations have been analyzed in terms of reliability. The network
configurations correspond to three representative networks already presented in section
4.2.

They are:

1) Urban medium voltage network (section 4.2.2.1);

2) Semi-urban medium voltage network (section 4.2.2.3);
3) Rural medium voltage network (section 4.2.2.4).

Parameter
Simulation step Value
s - speed; t - time
Regulatory threshold to consider long-duration
tmax,reg (min) 3
interruptions of supply

Response of the control centre for Fault toperator (min) 0.5

Detection, Isolation and Service Restoration
(FDIR) with smart grid solution t operator,j (min) 0.2

Response of maintenance crew t 1 (min) 14

Operation of load break switches in underground t 2 (min) 8

networks with secondary substations connected
in input-output s 3 (km/h) 37

Visual inspection of overhead lines s 4 (km/h) 6

Fault reparation t5 (min) 120

Table 5-2 Parameters of the reliability simulations

78
Figure 5-16 shows SAIFI as the percentage of tele-controlled switches in the networks
(referred to the total number of switches in the network) is increased. It is worth
mentioning that when there are 0% tele-controlled switches, there are still some manual
switches or breakers in the networks, so 0% tele-controlled switches is not equivalent to
0% automation. For each degree of automation and representative network, tele-
controlled switches have been placed regularly spaced along the feeders.

Figure 5-16 SAIFI as a function of the percentage of tele-controlled switches

As expected, the reliability of supply index corresponds to a higher improvement in the

urban two substations configuration than in the other two configurations considered
because of its topology. The worst result for the reliability of supply is obtained in the MV
rural network.

To perform this analysis two versions of the networks have been built, one with a low
degree of automation and one with a high degree of automation. The following table
shows the percentage of tele-controlled switches considered for each of these networks.

Network Low degree of High degree of

configuration automation automation
Two substations 0% 20%
Substation ring 0% 10%
MV Rural 0% 19%
Table 5-3 Percentage of tele-controlled switches in the networks

79
For a high degree of automation, the substation ring configuration has less tele-
controlled switches than the two substation configuration, because the substation ring
network refers to a semi-urban area.
However, in the MV rural network the percentage is high because it refers to the total
number of switches in the network, which is usually very low, as there are few loops in
that network, which makes the installation of a switch (manual or tele-controlled) less
useful. In Figure 5-16 and Figure 5-17 the rural network has only two points because
there are very few switches and loops in a rural network making meaningless
considering higher degrees of automation.
Figure 5-17 shows the SAIDI in the selected representative networks. In all the networks
SAIDI is below 60min. As with SAIFI, the lower degrees of automation achieve a
significant reduction of the reliability indexes, but there is a saturation effect, meaning
that the relative improvement of very high automation degrees is less significant. Such a
result can be useful to perform cost-benefit analysis for network automation
functionalities and technologies.

Figure 5-17 SAIDI as a function of the percentage of tele-controlled switches

80
6 Conclusions

This report presents the outcomes of the latest power distribution data gathering and
assessment exercise carried out by the JRC. It aims to provide policy decision makers
and other power system stakeholders with analyses and tools to better understand the
energy transition challenges faced by electricity distribution system operators in Europe.

It is also intended to support research activities in this field by providing reliable input
data and tools to carry out realistic simulations and comprehensive techno-economic
studies on the emerging developments in the power distribution system.

The electricity grid represents indeed a critical asset for our society. Given the rapidly
changing energy system, the scientific community needs to be able to rely on robust
models and realistic data to contribute to answer the questions society and policy
making are posing.

To the best of our knowledge, the report represents the most comprehensive data
collection exercise on European distribution systems published so far. Given the vast
number of Distribution System Operators active in Europe, the mapping effort was
limited to the bigger distribution system operators, i.e. the 190 Distribution System
Operators serving more than 100,000 customers and hence subjected to the unbundling
requirements of the EU Electricity Directive. Of these Distribution System Operators, 79
responded to the JRC survey.

Based upon the collected data:

 36 distribution system indicators were built, split in three categories: network

structure, network design and distributed generation.
 A total of 13 different representative distribution networks with different voltage
levels were built: 3 large scale geo-referenced networks (urban, semi-urban and
rural) and 10 feeder-type networks with common topologies.

These indicators and models help to shed light on the different characteristics of some of
the major European distribution networks. To get a sense of the potentialities of this
approach, some applications were presented and discussed.

Two large-scale representative networks (rural and urban) were selected to analyse the
impact of increasing levels of renewable energy sources penetration, photovoltaic panels
and wind farms in particular, on the technical performance of the grid. The impact on
network voltages and network overloads caused by these intermittent energy sources
has been then monetised by means of a penalty cost function.

The analyses have shown that the number and size of photovoltaic units, as well as their
connection/siting on the network are all relevant parameters which need to be carefully
taken into account. In fact, preliminary results highlight how limiting the size of the
generation units would mitigate voltage and congestion problems and maximise the
renewable penetration/network investment ratio. The case studies suggest that smaller
units spread throughout the network — compared to fewer bigger size (higher peak
capacity) units connected to it — would mitigate operation problems and would allow
deferring or avoiding future network reinforcements. A careful consideration of the local

81
conditions of each distribution area, as well as of the different connection patterns -
including unit sizes, technologies and location within the network - is of paramount
importance to minimise adverse impacts on system operation.

The installation of dispersed storage units combined with local photovoltaic plants
represents another way to mitigate the voltage spread introduced by the increasing
penetration of photovoltaics connected to the distribution network. However, the
analysis carried out through the representative distribution networks indicates that a
voltage spread reduction is only observed when big storage units are installed in
combination with each installed photovoltaic unit. The relatively high cost of batteries
suggests that other solutions should be considered to mitigate voltage spread in the
distribution network (e.g. inverters).

Finally, a reliability analysis showed how the System Average Interruption Frequency
Index might be improved (reduced) by increasing the installation of tele-controlled
switches in the distribution network, that is, by increasing the level of automation of the
considered distribution grid.

Representative distribution networks proved to be a valuable tool to simulate the impact

of increasing shares of renewable energy sources on network voltages and network
overloads. Representative distribution networks however, can also be useful in other
types of analyses, e.g. scalability and replicability analyses, cost-benefit analyses,
transmission and distribution systems inter-link and interdependencies analyses.

The JRC will continue to carry out its scientific and policy support activities in the power
system fields to better understand and address the challenges DSOs face in the
transition to a smarter energy system. In order to increase the knowledge base of the
evolving electricity distribution sector, the results of these activities will be made publicly
available23.

This report can be then seen as the first step of a periodic mapping and modelling
exercise, which the JRC aims to continue with the support of the relevant electricity
system stakeholders, in order to help understanding the merits, challenges and options
of the electricity system transition.

23
More information on the project and the representative networks in Matlab/Matpower format will be available
on: http://ses.jrc.ec.europa.eu/distribution-system-operators-observatory

82
Annex A: Indicator box plots

The following figures represent the box plot of the main used indicators as obtained from
the DSOs Observatory database. Additionally the values used for the built representative
distribution networks are shown. The circle represents the urban network, the star
represents the semi-urban network and the diamond represents the rural network. The
rest of the information refers to the values in the DSO Observatory database. In
particular, the red line is the median, the blue box represents the interval comprised
within the 0.25 and 0.75 percentiles, meaning that 50% of the DSOs have an indicator
value which is contained in the box. The black lines represent the full range, including
the minimum and the maximum values.

The ratio of the number of LV consumers per MV consumer has been set to the median
for the three types of areas, given that there was not information broken down per type
of area in the database.

Figure A- 1 LV consumers per MV consumer

Finland is an example of rural networks. In Finland the average is 0.07 LV km / (LV

Consumer), higher than the median. In the same way in the representative networks,
the more rural the area is, the higher the value of LV km / (LV Consumers).

83
 Figure A- 2 LV network length per LV consumer

In rural areas the LV underground ratio is lower than in semi-urban areas, and the LV
underground ratio in semi-urban areas is lower than in urban areas.

Figure A- 3 LV underground ratio (LV underground circuit length divided by total LV circuit length)

The number of LV consumers per MV/LV substation is higher in urban areas than in rural
areas. This is due to the higher density in the urban areas and the higher dispersion in
the rural areas. Typically in urban areas the constraint for connecting more consumers to
the MV/LV substation is power capacity, while in rural areas the constraint is the number
of nearby consumers.

84
 Figure A- 4 Number of LV consumers per MV/LV substation

The MV/LV transformer capacity per LV consumer is near the 0.75 percentile in the three
types of areas.

Figure A- 5 MV/LV transformer capacity per LV consumer

The MV circuit length per MV supply point is around 0.5 in The Netherlands, 0.65 in
Belgium and 0.95 in Finland. In the same way the MV circuit length per MV supply point
is higher in rural areas than in urban areas.

85
 Figure A- 6 MV network length per MV supply points

Typically underground cables are more common inside settlements than outside
settlements. Therefore, the MV underground ratio is lower in rural areas (which also
include MV feeders connecting several settlements) than in semi-urban areas, and it is
also lower in semi-urban areas than in urban areas. As the urban network represents the
center of a city, we have modeled its MV network as fully underground.

Figure A- 7 MV underground ratio (MV underground circuit length divided by total MV circuit
length)

86
In the three types of areas, the number of MV supply points per HV/MV substation is
near the median.

Figure A- 8 MV supply points per HV/MV substation.

87
Annex B: Other indicators

In this annex the remaining indicators not presented in the previous chapters are
reported for the interested reader.

Figure B- 1 LV consumers per area

Figure B- 2 Area per MV/LV substation

89
 Figure B- 3 Number of MV consumers per area

Figure B- 4 LV circuit length per area of distribution

90
 Figure B- 5 MV circuit length per area of distribution

Figure B- 6 Area covered per capacity of MV/LV substation

91
 Figure B- 7 Area per HV/MV substation

Figure B- 8 Capacity of HV/MV substation per MV supply point

92
Figure B- 9 Ratio of capacity of MV/LV substations per capacity of HV/MV substation

Figure B- 10 Area per capacity of HV/MV substations

93
Figure B- 11 HV circuit length per HV supply point

Figure B- 12 HV circuit length per area

94
 Figure B- 13 HV underground ratio

Figure B- 14 Number of electric vehicle public charging points per consumer

95
Figure B- 15 SAIDI for long unplanned interruptions

Figure B- 16 SAIFI for long unplanned interruptions

96
Figure B- 17 Typical transformation capacity of HV/MV substations

Figure B- 18 Total generation installed capacity per consumer

97
Figure B- 19 Percentage of generation connected to LV per technology

Figure B- 20 Percentage of generation connected to MV per technology

98
 Figure B- 21 Percentage of generation connected to HV per technology

Figure B- 22 Total Installed Capacity/consumer connected to LV, MV and HV

99
100
Annex C: On-line survey

1. Identification
Company/Association
Country
Phone
Fax
Email
Comments

2. General information (Distribution Business - Basic Figures, Structure &

Ownership)
A. Basic data
Legal Name of the DSO
Country
Regions and/or municipalities covered
Distributed Annual Energy (on average) (GWh)
Area of Distribution Activity (approximately) (km2)

B. Distribution business
Ownership of the DSO
A. Private
B Public state owned
C Public owned by municipality
D Other
Is the DSO part of a bigger group operating in the power industry?
If yes, type of unbundling with respect to the parent company:
Business in the power sector the company (or their group) operate besides
distribution (e.g. generation, transmission, supply/retail)

C. Customers
Total Number of Customers connected
Number of LV (< 1 kV) Customers
Number of MV (1- 36 kV) Customers
Number of HV (> 36 kV) Customers

D. Circuit length per voltage level (km)

Total
LV (< 1 kV)
of that Overhead
of that Underground
MV (1-36 kV)
of that Overhead
of that Underground
HV (> 36 kV)
of that Overhead
of that Underground

101
E. Technical data
Number of HV/MV Substations
Total installed capacity of HV/MV Substations (MVA)
Number of MV/LV Secondary Substations
Total installed capacity of MV/LV Secondary Substations (MVA)
Total installed capacity of generation connected (MW)
Installed capacity of generation connected to LV networks (MW)
Number of electric vehicle public charging points

F. Reliability
Reliability indexes (annual value of each reliability index for long unplanned
interruptions).

Reliability index Value LV** MV** HV**

SAIDI
(min./customer)
SAIFI (int./customer)

Please fill in the following table in case your reliability indexes are not the proposed
ones.

No. Reliability Unit Value LV** MV** HV**

Index*
1
2
3
4
5
6

G. Comments
Please mention here any comments or suggestions you may have

3. Further Collaboration

Are you interested in providing more customized information with the purpose of
building distribution networks representative of your company/country in a second phase
of this project?

4. Additional data
Would you like to provide additional data? If is it so, please choose among the three
categories (one or more):
□ Network structure
□ Distributed generation
□ Reliability
□ I can't provide additional data

102
Network structure

Network Data:
Typical transformation capacity of HV/MV Substations (MVA)
Typical transformation capacity of the MV/LV Secondary Substations in
urban areas (kVA)
Typical transformation capacity of the MV/LV Secondary Substations in rural
areas (kVA)
Average number of MV/LV Secondary substations per feeder in urban areas
Average number of MV/LV Secondary substations per feeder in rural areas
Average length per MV feeder in urban areas
Average length per MV feeder in rural areas
Number of TSO-DSO interconnection points
Voltage levels of the distribution networks (kV)
Typical number of voltage levels concatenated in distribution (for example 1
LV level, 1 MV levels and 1 HV level)
Degree of automation in the MV network [Type of smart grid automation
equipment and penetration]:
(e.g. Circuit breaker, Tele-controlled circuit breaker, Switch (on-load), Tele-
controlled switch, Fault detector, Directional fault detector, Recloser, …).

Substations equipped Degree of penetration Percentage of substations

with (low/medium/high)24 equipped with these
Monitoring/Automation equipment (%)
Equipment*
1
2
3
4
5
6
7
8
9
10

24
Low penetration is 0-5%, medium penetration is 5-20% and high penetration above 20%.

103
Distributed generation
Generation connected to distribution network (ONLY!)

Total Total Connected Connected Connected

Installed Gross to LV to MV to HV
Capacity Electricity (1kV) (1-36 kV) (>36kV)
[MW] Generation [%] [%] [%]
[GWh]
Photovoltaic
Wind
Biomass
Waste
Hydro

Reliability
Are the reliability indexes measured per type of area?
If yes, in what areas? What are the reliability indexes (annual value of each reliability
index per type of area, for long unplanned interruptions)?

Value
Urban-SAIDI (min./cust.)
Urban-SAIFI (int./cust.)
Rural-SAIDI (min./cust.)
Rural-SAIFI (int.cust.)

Please fill in the following table in case your reliability indexes or area type are not the
proposed ones.

Area type Reliability Index Units Value

Area 1
Area 2
Area 3
Area 4

Conclusion
Thank you for responding to our questionnaire. Are there any other questions that we
should have asked?
Do you wish to upload additional documents related to your activity? (maximum 1mb per
file; you can upload multiple files)

104
 Annex D: Cost functions

3.5

2.5
Cost (€/kWh)
2

1.5

0.5

0
0.7 0.8 0.9 1 1.1 1.2 1.3
Voltage (p.u.)

Figure D-1 Voltage cost function

Figure D-2 Overload cost function25

25
Sn is the nominal rating of the element in kVA.

105
106
References

CEER. Status Review on the Transposition of Unbundling Requirements for DSOs and Closed
Distribution System Operators. Brussels: Council of European Energy Regulators,
2013.

CEER. The Future Role of DSOs - A CEER Conclusions Paper. Brussels: Council of European
Energy Regulators, 2015.

Covrig, Catalin Felix, et al. Smart Grid Projects Outlook 2014. European Union, 2014.

DG Energy & Transport. "Note of DG Energy & Transport on directives 2003/54/EC and
2003/55/EC on the internal market in electricity and natural gas." 16.1.2004.

EnerNex. Eastern Wind Integration and transmission Study (Eastern Wind Dataset). EnerNex
Corporation, 2011.

EPIA. Connecting the sun. EPIA, 2012.

Eurelectric. Power Distribution in Europe – Facts & Figures. Brussels: Eurelectric, 2013.

Eurelectric. "The Role of Distribution System Operators (DSOs) as Information Hubs."

Brussels, 2010.

European Commission. Energy Markets in the European Union in 2011. Luxembourg:

European Union, 2012.

EWEA. Pure Energy: Wind energy targets for 2020 and 2030. EWEA, 2011.

Giordano, Vincenzo, et al. Smart Grid Projects in Europe: lessons learned and current
developments, 2012 update. Luxembourg: European Union, 2013.

Giordano, Vincenzo, Flavia Gangale, Gianluca Fulli, and Manuel Sanchez Jimenez. Smart Grid
Projects in Europe: lessons learned and current developments. Luxembourg:
European Union, 2011.

Glachant, Jean-Michel, Vincent Rious, and Jorge Vasconcelos. What future(s) for the EU
power transmission industry? European University Institute, 2015.

Gómez, T., Mateo, C., Sánchez, A., Frías, P., Cossent, R. "Reference network models: a
computational tool for planning and designing large-scale smart electricity
distribution grids." In High Performance Computing in Power and Energy Systems, by
Siddhartha Kumar Khaitan and Anshul Gupta. Berlin: Ed. Springer, 2013.

Gupta, Siddhartha Kumar Khaitan and Anshul. High Performance Computing in Power and
Energy Systems. Heidelberg New York Dordrecht London: Springer, 2013.

107
J. Wilkes, I. Pineda and G. Corbetta. Wind energy scenario for 2020. EWEA, 2014.

Jenkins, J. D. Economic regulation of Electricity Distribution Utilities under high penetration

of Ditributed Energy Resources: applying an incentive compatible menu of contracts,
reference network model and uncertainty mechanism. Boston: MIT, 2014.

Luo, X., Wang, J., Dooner, M., Clarke, J. "Overview of current development in electrical
energy storage technologies and the application potential in power system
operation." Applied Energy Vol. 137, 2015: 511-536.

Mateo, C., T. Gomez, A. Sanchez, J. Peco, and A. Candela. "A reference network model for
large-scale distribution planning with automatic street map generation." IEEE
Transactions on Power Systems. vol. 26, no. 1, 2011: pp. 190-197.

Mori, H., Iimura, Y. "Application of parallel tabu search to distribution network expansion."
IEEE Bologna, June 23-26, vol. 1, 2003: 6.

National Academies of Sciences, Engineering, and Medicine. Analytic Research Foundations

for the Next-Generation Electric Grid. Washington DC: The National Academies Press,
2016.

Pflugradt, N., Teuscher, J, Platzer, B., Schufft, W. "Analysing Low-Voltage Grids using a
Behaviour Based Load Profile Generator." International Conference on Renewable
Energies and Power Quality. Bilbao, 2013.

Poncela, Marta, Giuseppe Prettico, Nikoleta Andreadou, Miguel Olariaga Guardiola,

Gianluca Fulli, and Catalin Felix Covrig. Smart Grids Laboratories Inventory 2015.
Luxembourg: European Union, 2015.

Power Authority, Long Island. Transmission & Distribution Planning Criteria & Guidelines.
Long Island Power Authority, 2010.

R. Cossent, L. Olmos, T. Gómez, C. Mateo, P. Frías. "Distribution network costs under

different penetration levels of distributed generation." European Transactions on
Electrical Power. vol. 21, no. 6, 2011: pp. 1869-1888.

R. D. Zimmerman, C. E. Murillo-Sánchez, and R. J. Thomas. "MATPOWER: Steady-State

Operations, Planning and Analysis Tools for Power Systems Research and Education."
Power Systems, IEEE Transactions on, vol. 26, no. 1, 2011: 12 - 19.

Rigoni, V., Ochoa, L. F. Deliverable 3.7 "Characterisation of LV networks". Electricity North

West Limited (ENWL) - Low Voltage Network Solutions;, 2014.

Rodriguez, A., R. Cossent, P. Frías, and S. Izadkhast. "“gD3.5 Scalability & replicability rules.
Scaling up and replication rules of the GRID4EU demonstrations as inferred from the
SRA studies". GRID4EU Project Deliverable." 2016.

Román, J., Gómez, T., Muñoz, A., Peco, J. "Regulation of distribution network business." IEEE
Transactions on Power Delivery. vol. 14, no. 2, 1999: 662-669.

108
Ropenus, Stephanie, Sascha Thorsten Schröder, Henrik Klinge Jacobsen, Luis Olmos, Tomas
Gomez, and Rafael Cossent. Assessment of interactions between the economics of
distributed generators, distribution system operators and markets. Petten, The
Netherland: ECN, 2009.

Rudnick, H., Donoso, J.A. "Integration of price cap and yardstick competition schemes in
electrical distribution regulation." IEEE Transactions on Power Systems 4 , 2000:
1428–1433.

Ruester, Sophia, Ignacio Perez-Arriaga, Sebastian Schwenen, Carlos Batlle, and Jean-Michel
Glachant. From Distribution Networks to Smart Distribution Systems: Rethinking the
Regulation of European Electricity DSOs. European University Institute, 2013.

V. Giordano, I. Onyeji, G. Fulli, M. Sánchez Jiménez, C. Filiou. Guidelines for conducting a

cost-benefit analysis of Smart Grid projects. JRC Report, 2012.

van den Oosterkamp, Paul, et al. "The role of DSOs in a Smart Grid environment." 2014.

109
List of abbreviations and definitions
Benchmarking is the process of measuring an organization's internal processes then
identifying, understanding, and adapting outstanding practices from other organizations
considered to be best-in-class.
Circuit is a conductor or system of conductors through which an electric current is
intended to flow.
Circuit breaker is an automatic switch that stops the flow of electric current in a
suddenly overloaded or otherwise abnormally stressed electric circuit.
Demand response are changes in electric usage by end-use consumers from their
normal load patterns in response to changes in electricity prices and/or incentive
payments designed to adjust electricity usage, or in response to the acceptance of the
consumer’s bid, including through aggregation.
DSO is the abbreviation for distribution system operator.
Electric vehicle is an automobile that is powered entirely or partially by electricity.
Distributed generation is a method of generating electricity from multiple small
energy sources very near to where the electricity is actually used. Directive 2009/72/EC
defined distributed generation as “generation plants connected to the distribution
system”.
Fault detector is a device able to identify a defect in an electrical circuit due to which
the current is diverted from the intended path.
Feeder is a voltage power line transferring power from a distribution substation to some
point at which the power is broken into smaller circuits.
Fuse is a safety device that protects an electric circuit from excessive current, consisting
of or containing a metal element that melts when current exceeds a specific amperage,
thereby opening the circuit.
Low voltage network, in this document, is assumed to be the distribution network
installations working with a nominal voltage lower than 1kV.
LV is the abbreviation for low voltage.
Median is the middle number in a given sequence of numbers, taken as the average of
the two middle numbers when the sequence has an even number of numbers.
Medium voltage network, in this document, is assumed to be the distribution network
installations working with a nominal voltage in the 1..36kV range.
Microgrid is a group of interconnected loads and distributed energy resources (such as
distributed generators, storage devices, or controllable loads) within clearly defined
electrical boundaries that acts as a single controllable entity with respect to the grid. A
microgrid can connect and disconnect from the grid to enable it to operate in both grid-
connected or island-mode.
MV is the abbreviation for medium voltage.
MV/LV substation is a power station where the voltage is stepped down from medium
voltage to low voltage.
MV Supply points are the installations and loads directly supplied by the medium
voltage network, i.e. MV/LV substations and MV consumers.
High voltage network, in this document, is assumed to be the distribution network
installations working with a nominal voltage above 36kV. It is also commonly called sub-
transmission network in the literature.

111
HV is the abbreviation for high voltage.
HV/MV substation is a power station where the voltage is stepped down from high
voltage to medium voltage.
HV Supply points are the installations and loads directly supplied by the high voltage
network, i.e. HV/MV substations and HV consumers.
Quality of service is the degree to which the performances of the elements of the
electrical system result in power being delivered to consumers within accepted standards
and in the amount desired. It can be classified in commercial quality, continuity of
supply and waveform power quality.
Plug-in electric vehicle is any motor vehicle that can be recharged from an external
source of electricity, and the electricity stored in the rechargeable battery packs drives
or contributes to drive the wheels.
Recloser is a circuit breaker designed to trip when a fault is detected, then
automatically reclose after a set amount of time in an attempt to clear transient faults on
the feeder.
Reference Network Models are large scale distribution planning models, used in the
regulation for estimating the efficient cost required for building the distribution networks.
Revenue regulation is a form of price control applied to companies that are considered
to be regulated monopolies. Revenue regulation is designed to motivate regulated
companies to increase their efficiency whilst ensuring their economic viability.
SAIDI is the abbreviation for System Average Interruption Duration Index. It is a
measure of the duration of the interruptions.
SAIFI is the abbreviation for System Average Interruption Frequency Index. It is a
measure of the frequency of the interruptions.
Simultaneity factor is a ratio used to calculate the peak of a higher voltage installation
based on the peak of its lower voltage installations. It is required because the peaks of
the installations do not occur all at the same time and therefore when they are added
(e.g. to design an upstream installation), the peak of the aggregated profile is lower
than the sum of the individual peaks.
Smart grid is an electricity network that can integrate in a cost efficient manner the
behaviour and actions of all users connected to it - generators, consumers and those
that do both - in order to ensure economically efficient, sustainable power system with
low losses and high levels of quality and security of supply and safety.
Smart Meters are electronic measurement devices used by utilities to communicate
information for billing customers and operating their electric systems. Smart meters
enable two-way communication between the meter and the central system. Unlike home
energy monitors, smart meters can gather data for remote reporting.
Storage is a set of technologies capable of storing previously generated electric energy
and releasing that energy at a later time. Electrical Energy Storage (EES) technologies
may store electrical energy as potential, kinetic, chemical, or thermal energy, and
include various types of batteries, flywheels, electrochemical capacitors, compressed air
storage, thermal storage devices and pumped hydroelectric power.
Substation capacity is the amount of power that the substation can transform. It can
be expressed in kVA (kilovolt-amperes) or MVA (megavolt-amperes).
Switch is a device used to break or open an electric circuit or to divert current from one
conductor to another.
Transformer is a device that converts an alternating (A/C) current of a certain voltage
to an alternating current of different voltage, without change of frequency.

112
Underground ratio is defined as the percentage of the underground circuit length
referred to the total circuit length (overhead and underground) of the respective voltage
level.

113
List of figures
Figure 3-1 Distribution of connected customers ....................................................... 16
Figure 3-2 Distribution of yearly distributed energy ................................................. 16
Figure 3-3 Customers coverage per country ........................................................... 18
Figure 3-4 LV consumers per MV consumer ............................................................ 22
Figure 3-5 LV circuit length per LV consumer .......................................................... 22
Figure 3-6 LV underground ratio ........................................................................... 23
Figure 3-7 Number of LV consumers per MV/LV substation ....................................... 24
Figure 3-8 Transformers capacity per LV consumer ................................................. 24
Figure 3-9 MV circuit length per MV supply point ..................................................... 25
Figure 3-10 MV underground ratio ......................................................................... 26
Figure 3-11 Number of MV supply point per HV/MV substation .................................. 27
Figure 3-12 Typical transformation capacity of the MV/LV secondary substations in urban
areas (kVA) ........................................................................................................ 27
Figure 3-13 Typical transformation capacity of the MV/LV Secondary Substations in rural
areas ................................................................................................................. 28
Figure 4-1 Schematic view of the methodology used to build the representative
distribution networks ........................................................................................... 36
Figure 4-2 Street map image processing: identification of streets in the urban network 37
Figure 4-3 Street map image processing: building location in the urban network ......... 37
Figure 4-4 Street map image processing: semi-urban network .................................. 38
Figure 4-5 Street map image processing: rural network ........................................... 38
Figure 4-6 Urban network .................................................................................... 40
Figure 4-7 Urban MV feeders ................................................................................ 42
Figure 4-8 Semi-urban network ............................................................................ 43
Figure 4-9 Semi-urban MV feeders ........................................................................ 45
Figure 4-10 Rural network .................................................................................... 46
Figure 4-11 Rural MV feeders................................................................................ 48
Figure 4-12 Typical transformation capacity of the MV/LV secondary substations in urban
areas ................................................................................................................. 49
Figure 4-13 Typical transformation capacity of the MV/LV secondary substations in rural
areas ................................................................................................................. 49
Figure 4-14 Average number of MV/LV substations per feeder in urban areas ............. 50
Figure 4-15 Average number of MV/LV substations per feeder in rural areas ............... 50
Figure 4-16 Average length per MV feeder in urban areas ........................................ 51
Figure 4-17 Average length per MV feeder in rural areas .......................................... 51
Figure 4-18 Two substation urban network ............................................................. 52
Figure 4-19 MV/LV Transformer capacity ................................................................ 53

115
Figure 4-20 Urban switching station ...................................................................... 54
Figure 4-21 MV/LV Transformer capacity ................................................................ 55
Figure 4-22 Semi-urban substation ring ................................................................. 55
Figure 4-23 MV/LV Transformer capacity ................................................................ 56
Figure 4-24 Rural MV network............................................................................... 57
Figure 4-25 MV/LV Transformer capacity ................................................................ 58
Figure 4-26 Urban LV network .............................................................................. 59
Figure 4-27 Semi-urban LV network ...................................................................... 60
Figure 5-1 Expected solar PV penetration in 2020 and 2030 (EPIA) ........................... 65
Figure 5-2 RES penetration scenarios .................................................................... 67
Figure 5-3 Bus voltages in the rural network for a 100% PV size limit (Peak PV hours) 68
Figure 5-4 Bus voltages in the rural network for a 100% PV size limit (Peak demand
hours) ............................................................................................................... 69
Figure 5-5 Voltage spread in the rural network ....................................................... 69
Figure 5-6 Voltage cost impact on the rural network ................................................ 70
Figure 5-7 Voltage map of the rural network at 14:00 in Sc. 0 .................................. 71
Figure 5-8 Voltage map of the rural network at 14:00 in Sc. 2, with a 200% onsite PV
size limit ............................................................................................................ 71
Figure 5-9 Aggregated profiles in year 2030 ........................................................... 72
Figure 5-10 Number of overloads in the rural network ............................................. 74
Figure 5-11 Network overload costs in the rural network .......................................... 74
Figure 5-12 Number of overloads (branch*hour) ..................................................... 75
Figure 5-13 Voltage spread in Sc. 1 with increased storage unit capacity ................... 76
Figure 5-14 Voltage and overload cost in Sc. 1, for a 100% onsite PV size limit, with
increased storage unit capacity ............................................................................. 76
Figure 5-15 Parameters used in the simulations to obtain SAIDI and SAIFI ................ 78
Figure 5-16 SAIFI as a function of the percentage of tele-controlled switches ............. 79
Figure 5-17 SAIDI as a function of the percentage of tele-controlled switches ............. 80

116
List of tables
Table 2-1 DSOs number per Country (Eurelectric 2013) ........................................... 10
Table 3-1 Participation per country ........................................................................ 17
Table 3-2 Network structure and reliability indicators ............................................... 19
Table 3-3 Network design indicators ...................................................................... 20
Table 3-4 Distributed generation indicators ............................................................ 20
Table 3-5 Subset of the total DSOs indicators used to build the large-scale
representative distribution networks ...................................................................... 21
Table 4-1 Representative networks ........................................................................ 33
Table 4-2 DSO Database network indicators ........................................................... 39
Table 4-3 Indicator name and representative network ratios ..................................... 39
Table 4-4 Low voltage feeders .............................................................................. 61
Table 4-5 Medium to low voltage transformers ........................................................ 62
Table 4-6 Medium voltage feeders ......................................................................... 62
Table 4-7 High to medium voltage substations ........................................................ 62
Table 5-1 Percentage of consumers installing PV for each scenario and onsite PV size
limit .................................................................................................................. 72
Table 5-2 Parameters of the reliability simulations ................................................... 78
Table 5-3 Percentage of tele-controlled switches in the networks .............................. 79

117
 How to obtain EU publications

Our publications are available from EU Bookshop (http://bookshop.europa.eu),

where you can place an order with the sales agent of your choice.

The Publications Office has a worldwide network of sales agents.

You can obtain their contact details by sending a fax to (352) 29 29-42758.

Europe Direct is a service to help you find answers to your questions about the European Union
Free phone number (*): 00 800 6 7 8 9 10 11
(*) Certain mobile telephone operators do not allow access to 00 800 numbers or these calls may be billed.

A great deal of additional information on the European Union is available on the Internet.
It can be accessed through the Europa server http://europa.eu
 LD-NA-27927-EN-N
JRC Mission

As the Commission’s
in-house science service,
the Joint Research Centre’s
mission is to provide EU
policies with independent,
evidence-based scientific
and technical support
throughout the whole
policy cycle.

Working in close
cooperation with policy
Directorates-General,
the JRC addresses key
societal challenges while
stimulating innovation
through developing
new methods, tools
and standards, and sharing
its know-how with
the Member States,
the scientific community
and international partners.

Serving society
Stimulating innovation
Supporting legislation

doi: 10.2790/701791

ISBN 978-92-79-58416-9

120